EPA-600/2-75-016
August 1975 Environmental Protection Technology Series
INCINERATOR OVERFIRE
MIXING DEMONSTRATION
U.S. Environmental Protection Agency
Office of Research and Development
Washington, D.C. 20460
-------
EPA-600/2-75-016
INCINERATOR OVERFIRE
MIXING DEMONSTRATION
by
T. J. Lamb, R. H. Stephens, C. M. Mohr,
P. C. Levins, L. K. Fox, and A. F. Sarofim
Arthur D. Little, Inc.
20 Acorn Park
Cambridge, Massachusetts 02140
Contract No. 68-02-0204
ROAP No. 21AUZ-015
Program Element No. 1AB015
EPA Project Officers:
James D. Kilgroe
Industrial Environmental Research Laboratory
Research Triangle Park, North Carolina 27711
and
Donald A. Oberacker
Municipal Environmental Research Laboratory
Cincinnati, Ohio 45268
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, D.C. 20460
August 1975
-------
TABLE OF CONTENTS
Page
List of Figures v
List of Tables vii
Nomenclature ix
SECTION I - CONCLUSIONS 1
REDUCTION OF COMBUSTIBLE EMISSIONS 1
JET MIXING 1
FURNACE FLUID FLOW 2
BED BURNING 2
SECTION II - RECOMMENDATIONS 3
SECTION III - INTRODUCTION 5
BACKGROUND 5
OBJECTIVES 6
APPROACH AND SCOPE 7
SECTION IV - SUMMARY 9
TESTING PROGRAM 9
VERIFICATION OF MODELS 9
STRATEGY FOR EMISSION CONTROL 11
TEMPERATURE CONTROL 13
DESIGN GUIDELINES 14
SECTION V - CONCEPTS IN INCINERATION 17
STRATEGY FOR EMISSION CONTROL 17
EFFECT OF MIXING 18
SECTION VI - EXPERIMENTAL PROGRAM 45
TEST INCINERATOR 45
INCINERATOR MODIFICATIONS 49
CONDUCT OF TESTS 50
ANALYTICAL EQUIPMENT 59
DATA REDUCTION 69
111
-------
TABLE OF CONTENTS (Continued)
SECTION VII - TEST RESULTS
MACROSCALE MIXING MODEL
FLU ID-FLOW PATTERNS
MICROSCALE MIXING MODEL
RESULTS OF PARTICULATE TESTS
SECTION VIM- DESIGN GUIDELINES
TEMPERATURE
FURNACE DESIGN
UNDERFIRE AIR DISTRIBUTION AND FLOW RATE
OVERFIRE AIR JETS
APPENDIX A
APPENDIX B
APPENDIX C
APPENDIX D
APPENDIX E
APPENDIX F
APPENDIX G
TEST MEASUREMENT AVERAGES
CALCULATION OF UNMIXEDNESS FACTOR OF REFUSE
BEDOFFGAS
MATHEMATICAL CHARACTERIZATION OF QUENCHING
BASIC PRINCIPLES OF COMBUSTION
CHARACTERIZATION OF THE BED-BURNING PROCESS
DERIVATION OF THE UNMIXEDNESS FACTOR
CONVERSION FACTORS FOR SI UNITS
Page
75
75
76
77
87
93
93
93
95
98
103
109
113
121
131
145
147
IV
-------
LIST OF FIGURES
Figure No.
1
2
3
4
5
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Emission Factors vs.Temperature
Heterogeneous Character of Flue Gases
The Unmixedness Factor ^/c^ in Turbulent Flames
Concentration Fluctuations
Effect of Stoichiometry on Emissions for a Fixed Turbulence
Level
Effect of Turbulence Level on Emissions for a Fixed
Stoichiometry
Overall Effect of Micromixing on Emissions
Offgas Composition of C0/C02
Theoretical Relationship
Dependence of Emission Factor on Temperature and Unmixedness
Factor
Theoretical Relationship
Dependence of Emission Factor on Temperature and Unmixedness
Factor (Revised Plot of Figure 10)
Cross Section of Newton 500-TPD Incinerator Showing
Sampling Locations
Overall Emissions Measurements Approach
Sampling System
Gas Analysis System
Static Pressure and Velocity (AP) System
Water-Cooled Sampling Probe
Thermocouple Radiation Shield
Gas Sampling Lines, Pumps and Meters
Recording Temperature and Pressure Systems
NDIR and FID Analyzers and Recorders
Gas Chromatograph and NDIR Analyzers
C02 vs. Temperature
Penetration vs. Jet Velocity
Confirmation of Bernoulli's Equation
(1-x) vs. (0t)b
Page
12
24
25
26
28
28
30
32
34
35
36
41
46
56
56
58
60
62
63
65
66
67
68
74
76
78
86
-------
LIST OF TABLES
Table No. Page
1 Composition of Bed Offgas 32
2 Summary of Calculated Terms 43
3 Velocity and Static Pressure Measurement Locations 59
4 Carbon Distribution Profile 70
5 Hydrogen Balance for Typical Run 72
6 Oxygen Balance in Typical Run 72
7 Summary of Unmixedness Factors 8^
8 Summary of Test Results Over Bed 82
9 Comparison of Estimates for (1-x) 8"*
10 Calculation Results Using Different Values for x °*
11 Relative Agreement Between L/C**) t AND (v/c^)taic 86
12 Sampling Points Per Radius eXP 88
13 Gaseous Pollutant Emissions 88
14 Stack Particulate Data 89
15 Particulate Analyses 89
16 Trace Metals Analysis of Filter Catch ^
17 Trace Metals Analysis of Probe and Cyclone Catch "'
18 Excess Air Requirements for Efficient Combustion 9^
19 Theoretical Air Requirements of Refuse and Other Fuels ^6
vn
-------
NOMENCLATURE
A Area
Ac Area of crossflow, Eqs. (5) and (14)
Af Area of flow above jets
A, Area of overfire jet
A0 Refuse bed area
C Concentration of off gas expressed as mol/volume or as mol/fraction
C Mean concentration
C Concentration of off gas at onset of quenching
Cs Concentration of off gas in stoichiometric mixture
CT Total molar concentration, mols/volume e.g. (P/RT)
Cp Heat capacity
Cp Cp Heat capacity of combustion products of stoichiometric mixture, of excess air
s, xs
c' Concentration fluctuations about the mean
c'2 Unmixedness factor (root mean square of c')
(v c^bo Unmixedness factor for tests without jets with breech, over bed
(v/c'2); (\/c'2): Unmixedness factor for tests with jets with breech, over bed
J> Jo
Du Diffusivity of fuel in air, gm/cm2 /sec, etc.
dj Diameter of jet
Ea Eaf Ear Activation energy, for forward reaction, for reverse reaction
f(0) VC^/AC defined by Figure 3
g Acceleration of gravity
gc Conversion factor of weight to mass
HHV Higher heating value
K Equilibrium constant
K^, Equilibrium constant for water-gas shift
K E Kinetic energy
k Constant or thermal conductivity (App. C)
k3 Chemical kinetic constant of reaction (3), for example
L Characteristic dimension of system
Lp Penetration distance of jet
m Mass
IX
-------
NOMENCLATURE (Continued)
N - Number of jets
n Moles per mole of off gas
nc Moles of combustion products per mole of off gas
no Moles of oxygen required to bar one mole of off gas
nH o>nN > nco Moles of water vapor per mole of off gas, of nitrogen, of C02, etc.
ns Total moles in stoichiometric mixture containing one mole of off gas
nT Total moles in mixture containing one mole of off gas
nxs Moles of excess air (not oxygen) per mole of off gas
(O2 Req)b, (O2 Req)bo Oxygen required to complete combustion of off gas without jets in
breech, initial
(O2 Req)j Oxygen required to complete combustion of off gas in breech'
with jets on
(O2 Req)jo Oxygen required to complete combustion of off gas at top of
bed with jets on
P Pressure
Pa Partial pressure of component a, for example
P0- P Differential pressure between grate and breech
Q Flow rate, volume/time
Qc Flow rate of cross flow to jet axis
Qf Flow rate of gases after jets or exit the breech
Q. Flow rate of jets
Q0 Flow rate of off gas from refuse bed
R Ideal gas constant
S Jet spacing parameter
T Temperature
T Adiabatic flame temperature
3
Tc Temperature of gas in cold zone of incinerator
TH Temperature of gas in hot zone of incinerator
T0 Ambient temperature
T- Jet temperature
Tc Temperature of crossflow gases
Tf Temperature of gases after jets
-------
NOMENCLATURE (Continued)
t Time
u Velocity
ur ,u Cold zone velocity, initial value
*- co
uc Crossflow velocity (before jets)
uf Furnace velocity (after jets)
UH UH Hot zone velocity, initial value
Uj Jet velocity
u Initial velocity
o
V Volume of gas
Vf Furnace volume
x Moles fraction of off gas in mixture of off gas plus jet air
y Mole fraction of off gas
yc Mole fraction of off gas in crossflow (before jets) gases
yf Mole fraction of off gas in furnace (after jets) gases
ys Mole fraction of off gas in stoichiometric mixture
Ayb (y - ys) evaluated for base tests (no jets)
Ay. (y - yj evaluated for jet tests
j »
z Distance
zo Height at top of refuse bed
zc Height to cold zone in breech
ZH Height to hot zone in breech
z Distance from initial surface to quench plane (App. C)
zs Distance from initial surface to stoichiometric plane (App. C)
z' Distance from initial surface to zero-oxygen plane (App. C)
XI
-------
Greek Symbols
a thermal diffusivity or constant [Eq. (F-4)]
(3 mixing energy constant [Eq. (22)]
7 ratio of furnace height to width
e rate of input of mixing energy per mass of system
77 ratio of furnace length to width, also constant [Eq. (C-12)]
p density
pc density of cold zone gases
PH density of hot zone gases
pc density of crossflow gases
pf density of furnace gases (after jet)
Pj density of jet air
T residence time
0 available oxygen/oxygen required to complete combustion
0b ^b ^ f°r base tests (no jets), initial value at top of refuse bed
0 0 0 for jet tests, initial value at top of refuse bed
j. JG
Common Subscripts
b baseline (no jets)
C cold zone
c crossflow (before jets)
f furnace (after jets)
H hot zone
j Jet
o initial, usually refers to top of refuse bed
q quenched conditions
s stoichiometric conditions
Gas Compositions
(% CO2) percent CO2 in gas mixture
(% CO2 )b percent CO2 without jets on
(% CO2 )j percent CO2 with jets on
% CO2 ls percent CO2 in stoichiometric mixture
CO carbon monoxide (ppm)
CO|S carbon monoxide in stoichiometric mixture with defined CO/CO2
% O2 percent oxygen
Xll
-------
SECTION I
CONCLUSIONS
Overfired air jets have been used in coal stokers to induce turbulence for increased
combustion efficiency and reduced emissions of combustible pollutants such as CO, H2,
hydrocarbons, and soot. A mixing system of this type was installed in an existing municipal
incinerator under Contract 68-02-0204 with the Environmental Protection Agency. Tests
were conducted to determine the effects of and design criteria for mixing jets. As a result of
the testing program and subsequent analyses, Arthur D. Little, Inc., reached the following
conclusions:
REDUCTION OF COMBUSTIBLE EMISSIONS
1. Combustible emissions can occur either by quenching where the temperature
of the gases is too low to sustain combustion, or by a mixing limitation
where the gases are not completely mixed with the available oxygen, and
thus will not burn out in the furnace within the residence time. The former
becomes significant when gas temperatures are below about 1400°F; the
latter is significant when gas temperatures are above about 1800°F. By
adjusting air input to maintain furnace temperature within the range of
1400 to 1800°F, both quenching and mixing limitation were avoided.
2. Overfire jets reduced combustible emissions resulting from the mixing limita-
tion. This reduction is attributed to the greater availability of oxygen rather
than to large increases in mixing intensity.
A theory based on statistical descriptions of the turbulent mixing was
developed and verified by test data. This theory provides a quantitative
expression for the heretofore qualitative concepts of the three T's - time,
temperature, and turbulence. This theory can be used to show under what
conditions combustible emissions caused by inadequate mixing can be
reduced by increasing mixing intensity through the use of sidewall jets or by
increasing furnace residence time.
3. Overfire jets were ineffective in reducing combustible emissions due to
quenching. These emissions are believed to be more amenable to control
through proper furnace design and underfire air control than through mix-
ing. Unfortunately, the test incinerator was not sufficiently versatile to allow
us to explore this emission cause in depth.
JET MIXING
1. Test results showed that the penetration of a sidewall jet into the gases from
the refuse bed can be predicted using the design equations presented in this
report.
1
-------
2. Sidewall overfire jets were found to be very effective in controlling the
temperature of the overfire region.
3. The jet test configurations, such as opposed jets, interlaced jets, or jets on
one side, were shown to be equally effective for attaining adequate uniform
mixing and temperature control.
4. While the jets did not have a significant effect on increasing the mixing
intensity in the test furnace, extrapolations of the theory, which was verified
in the test furnace, suggest that such an effect could be significant in smaller
furnaces.
FURNACE FLUID FLOW
1. The characteristics of the stratified layers, including their acceleration due to
buoyancy forces, can be predicted from the geometry of the furnaces and
temperatures along the refuse bed surface.
2. When the gases exit directly over the cold end of the furnace (the burnout
grate) so that no forced mixing of the hot gases and cold gases occurs, then
the gases will stratify, forming stable layers. The colder gases are on the
bottom and quench combustion, while the hotter gases are on top and are
often oxygen-deficient. Both conditions can result in high combustible
emissions.
BED BURNING
1. The processes of drying, ignition, pyrolysis, combustion, and burnout occur
simultaneously throughout the refuse bed as witnessed by a constancy of the
ratio H2 O/CO2 throughout the furnace.
2. The composition of gases both leaving the refuse bed and also in oxygen-
deficient (unmixed) pockets is controlled by the water/gas shift equilibrium,
as confirmed by test data in both the breech of the test furnace and also
directly over the refuse bed.
3. In a furnace that is unsealed, or which has a large amount of uncontrolled
overfire air, the oxygen required for combustion is entrained into the refuse
bed from the overfire air space. Even when the underfire air is considerably
less than stoichiometric, the burning rate of refuse tends to be constant at
approximately 60 lb/hr/ft2.
-------
SECTION II
RECOMMENDATIONS
Based on the data collected in this experimental program, Arthur D. Little, Inc.,
recommends the following:
1. The use of overfire air jets to reduce combustible emissions from existing
municipal incinerators.
2. The use of overfire jets in small furnaces to increase mixing intensity. Design
criteria presented in this report should be utilized.
3. The realization of uniform mixing throughout the furnace through proper
control of underfire air and certain changes in furnace configuration. For
example:
a. To achieve uniform mixing with underfire air, a high-pressure drop
grate should be used, and the furnace should be sealed to prevent any
leakage which would tend to negate the effect of controlled air rates.
b. The furnace should be designed so that hot and cold gases are forced to
mix before exiting the furnace. This represents the most effective
technique for attaining a uniform front-to-back mixing.
c. Overfire jets should be designed to effect uniform side-to-side mixing,
as shown in this report.
-------
SECTION III
INTRODUCTION
BACKGROUND
The emissions from municipal incinerators have been the subject of considerable
discussion over the past decade in conjunction with growing public awareness of the adverse
effects of air pollution. The incinerator emission receiving the most attention to date is fly
ash, a particulate emission having a characteristic dimension between 0.1 and 100
microns (M). Approximately 70% (by weight) of fly ash has a characteristic dimension
greater than 5ju and is therefore easily removed using relatively simple devices. The smaller
particles are much more difficult to remove, however, but in recent years several advances
have been made in the state-of-the-art of small particulate removal. The development of
such devices as electrostatic precipitators and high-energy wet scrubbers both high-
efficiency collection devices has facilitated the removal of particles down to 0.5/u so that
99% of the particulate emission can be controlled. As municipal incinerators across the
country incorporate these types of emission controls, the prevalence of fly ash as a pollutant
will diminish.
On the other hand, in its report [ 1 ] "Systems Study of Air Pollution from Municipal
Incinerators" (Contract CPA-22-69-23 and dated March 1970). Arthur D. Little, Inc. (ADL)
stated that the municipal incinerator was a large source of such pollutants as carbon
monoxide, hydrocarbons and soot (small carbon particles). Unfortunately, these emissions
are not so easily controlled as is the fly ash. The soot particles are so small, in fact, that a
significant fraction will pass through even the highest efficiency control devices. The gaseous
emissions are relatively unaffected by particulate control devices. However, these emissions
are all products of incomplete combustion and, therefore, can be controlled by improving
combustion efficiency rather than by the use of expensive air pollution control devices. In
its report ADL recommended that methods be developed to improve the combustion
efficiency of municipal incinerators to preclude incomplete combustion.
Designers have long recognized that combustible emissions could be reduced by
increasing the combustion efficiency of a process. In fact, they have taken several steps to
maximize residence time and increase the amount of mixing within the furnace, but most
current design practice is based upon empirical rules or qualitative guidelines. Unfortu-
nately, many accepted engineering principles of combustion have not as yet been applied to
incinerators.
In February 1972, ADL published another report entitled, "Incinerator Overfire Mixing
Study" (Contract EHSD 71-6) [2] .This study,which reviewed the present understanding of
the combustion process in solid fuel equipment and municipal incinerators, postulated a
technique for improving combustion efficiency in which jets would be used to promote
mixing. A test program was proposed to demonstrate the use of overfire air and steam jets
to reduce the combustible pollutants being emitted from an existing municipal incinerator.
-------
In this previous study, ADL postulated that combustion in a municipal incinerator
occurred in three distinct zones an ignition-limited zone, a combustion-limited zone, and
a fuel-limited (burnout) zone. Under conditions normally found in incinerators, the oxygen
concentrations in the first and last zones are high, reflecting the low oxygen demand
required to sustain combustion. The central zone is oxygen-limited and has an extremely
low residual oxygen concentration. Similarly, the temperature in the central zone is
considerably higher than in either of the other two zones and the gas emanating from this
zone has a much greater concentration of CO and other combustible emissions. Unfortu-
nately, the oxygen-rich zones do not mix well with the oxygen-deficient central zone. The
model developed to determine gas-flow patterns within the incinerator indicated the zones
remain stratified long after the combustion process had been quenched. We, therefore,
suggested two approaches to eliminate the emissions resulting from incomplete combustion:
1. Distribution of underfire air along the grates in proportion to the oxygen
demand, and
2. The use of air and/or steam jets to induce turbulence, thus promoting greater
combustion efficiency.
To verify the model for burning, mixing, and stratification and to demonstrate the
effectiveness of overfire mixing to reduce combustible emissions, EPA awarded a contract to
ADL to test an existing municipal incinerator and to develop design and operating guidelines
which could be used for improving the combustion efficiency of other incinerators.
OBJECTIVES
Compared to other industrial sources of pollutants, the municipal incinerator is a
low-budget piece of equipment, often poorly operated or maintained. There has been little
incentive for either sophisticated testing or development work on the part of either the
municipality (the owner) or the engineer (designer). For that reason very few meaningful
data are available.
To extend the state-of-the-art, this program had several independent objectives. They
were to:
1. Develop a basic understanding of the incineration process by verifying the
three models presented in the previous report and extending these concepts
as far as possible to describe completely the combustion of refuse within an
incinerator;
2. Demonstrate the use of overfire jets on an existing municipal incinerator of
common design and showing their effect on combustible emissions;
-------
3. Develop guidelines for design and operation of the jet systems, because much
of the work is of a theoretical nature extending beyond the normal interests
of both the designer and the operator; and
4. Supply these data insofar as possible, because - independent of the focus of
the other objectives there is an extreme lack of basic operating and
emission data on incinerators.
APPROACH AND SCOPE
In developing design data for the use of overfire jets on municipal incinerators, we tried
to strike a balance between both the theoretical and the experimental approaches that are
required to satisfy the specific objectives of this program. Although we only tested a single
incinerator, our intent was to generalize the test results to other furnaces.
Full characterization of the incinerator would involve a complete input/output
analysis. Since the principal emphasis in the program was on combustible emissions, the
major data-collection effort was restricted to this area. Limited data on material input were
obtained to guide the overall data analysis. The information obtained during each of the
tests included:
Input: refuse tonnage, grate speeds and air flow rate;
Output: temperature, flow (velocity), and composition.
Gaseous composition data were required to reflect both the overall combustion
performance of the incinerator and the variation in combustible emissions. The following
species were routinely analyzed in the breaching region (and over the burning bed during the
study of that region):
Overall composition: O2, CO2, N2, and H2 O;
Combustible emissions: CO, H2, and THC (total gaseous hydrocarbons).
In addition, a carbon monoxide analyzer was run continuously in the duct leading to
the base of the stack to reflect the long-term trend in changes of emissions.
The key factor in the analytical testing program was the development of a high-
temperature testing probe to sample in the 2000° F (or greater) conditions found both in the
breech and over the burning bed. Continued service from a test probe under these
conditions for more than one week of testing proved to be very difficult, but because of
modification in the testing procedure and improvements in the probe design, we believe that
the technique developed in this program will be generally applicable to all similar instances
of emission testing in high-temperature environments.
In conducting this program, the initial series of tests (the baseline tests) was used
primarily to verify the analytical approaches and to ensure that the test data were consistent
-------
internally and with the physical concepts of incineration. With each series of tests, material
and energy balances were completed and the results were analyzed to indicate potential
weaknesses in the test data or to point out areas where present understanding of combustion
could not correlate the experimental results. Because of this procedure, there were changes
made in the analytical approach with corresponding increases in the accuracy of the
experimental data. The material balances for the final jet tests indicated the magnitude of
error in the analytical data to be less than 10%.
Upon verification of the initial testing data, it became apparent that many of the
theories used to model the combustion were not wholly adequate with respect to the
characterization of mixing and combustible emission burnout. Additional work in the
modeling of combustion was carried out to augment the models presented in our previous
report. The emphasis was on increasing the sophistication of the models so that the physical
phenomena could be more readily deduced from the test data. The analysis of the bed
model took on much less importance as the work progressed. The bed model is introduced
only in the Summary to the report. The detailed discussion of the bed model is contained in
Appendix E.
We did not intend for this report to be an exhaustive compendium of combustion
mechanics, but only that the principles and experimental results pertinent to the burnout of
combustible emission be fully explored and documented. We have incorporated the test data
in as complete a form as possible (Appendix A), and we feel it will prove to be a valuable
resource to those workers in the field who would like to extend this work still further, or
who may require data on existing furnaces for experimental support of their own theories.
-------
SECTION IV
SUMMARY
The primary objective of this program was the reduction of combustible emissions
from existing municipal incinerators through the use of overfire air jets to induce mixing.
The objectives cited in the Introduction of this report represent the various aspects required
to study this particular problem adequately.
TESTING PROGRAM
To provide the data required to verify the burning and mixing models developed in the
Incinerator Overfire Mixing Study (2), we planned a test program at an existing municipal
incinerator. We chose the incinerator of the City of Newton, Massachusetts, as the test
vehicle, because it is similar in design to many incinerators in the United States, it could be
easily retrofitted with an air jet system, and the City of Newton was kind enough to make
the facility available.
In the initial tests, we recorded composition, temperature, and velocity data in the
breech of the furnace to check out the analytical procedures and provide information
needed to design the jet system.
Following the procedure outlined in the previous study, we designed and installed an
overfire jet mixing system. The system consisted of ten 4-inch jets on each side of the
furnace, a-forced draft fan, and the ducting and instrumentation necessary to control the
system.
The test program included tests both in the breech of the incinerator and also over the
refuse bed. The breech tests were required to verify the fluid flow model and also to
confirm or deny the effectiveness of overfire mixing jets in reducing combustible emissions.
The bed data were necessary for verifying the bed burning model.
VERIFICATION OF MODELS
In the previous study (2), ADL postulated three models necessary to describe incinera-
tion: l)a bed-burning model, 2) a fluid-flow model, and 3) a jet-mixing model. All three
were verified either as postulated or with the refinements added as a result of the testing
work. Each model is summarized below.
Bed-Burning Model
The bed-burning model was originally proposed to describe the overbed gas composi-
tion and also the burning rates for refuse beds. The model was previously postulated on the
basis of separate drying, ignition, pyrolysis, combustion and burnout planes, but the results
-------
of the testing program showed that the processes occurred simultaneously throughout the
refuse beds. However, this fact does not alter the concept that the gases leaving the refuse
bed in the absence of oxygen are controlled by the water gas shift equilibrium. In support of
this, gas samples taken directly over the refuse bed contained concentrations of CO, CO2,
H2, and H2 O in direct proportion to what would have been predicted from the equilibrium
relation. However, each sample also contained a significant amount of oxygen, although the
previous model was based upon the existence of a plane of zero oxygen within the refuse bed.
This particular phenomenon strongly suggested that the combustible gases were not mixed
with oxygen and remained unmixed even after the gases had left the furnace. As a result a
fourth model the microscale mixing theory was deemed necessary to describe com-
pletely the combustion phenomenon that was occurring within the furnace volume. This
model is described in a later section of this summary and also in Section V.
Based upon the empirically determined amounts of oxygen that can be entrained in the
refuse bed and also upon the water gas shift gas equilibrium, we were able to develop
estimates for overbed composition of gases. This information was required in the model
developed for microscale combustion.
An additional ramification of the oxygen entrainment was the fact that the bed-
burning rates tended to remain constant at approximately 60 lb/hr/ft2. Surprisingly enough,
the bed-burning rate was not very sensitive to the underfire air rate in that the air deficiency
was supplied from the overfire region. This would not be the case either when the amount
of oxygen was very limited, or when greater than stoichiometric amounts of underfire air
were present.
Fluid Flow Model
The acceleration of gases away from the refuse bed can be mathematically charac-
terized using Bernoulli's equation:
2g(z-zo)
(1)
In a furnace that is completely sealed, the appropriate Tc is the temperature in the cold
region of the furnace, most likely the temperature over the burnout grate. In an unsealed
furnace the appropriate Tc is the ambient air temperature. Because of the large amount of
leakage that was inherent in all of the test runs at the Newton incinerator, it was obvious
that the furnace was unsealed. Equation (1) was verified for a Tc of approximately 70°F.
Jet Mixing Model
In a previous study (2), ADL postulated that jets could be used to mix cold regions,
such as the region over the burnout grate, with the hotter regions of the furnace so that the
overall temperature of the furnace could be controlled. To design such a jet it was necessary
10
-------
to arrive at a relationship between the jet penetration distance and the parameters of the
system. The jet tests were able to verify the design equation suggested by Ivanov (3):
(2)
In designing a jet system, the designer has the freedom to adjust the jet diameter location
and the jet velocity, with the only constraint being that the jet system not be allowed to
deliver an amount of air to the furnace that would quench combustion. The other variable
in the jet design equation is uc , the cross-flow velocity for the gases rising off the refuse bed
perpendicular to the row of the jets. This is determined from Eq.(l) above.
STRATEGY FOR EMISSION CONTROL
Origin of Combustible Emissions
Two phenomena which result in combustible emissions were identified in the course of
this study. The experimentally determined CO emission factor for the furnace is a function
of local furnace temperature as shown in Figure 1. In regions of the furnace where the
temperature was below approximately 1 100°F, emissions occurred because the reaction had
been quenched, i.e., it was too cold to sustain combustion.
On the other hand, when the temperature was high (above 1800°F), the corresponding
amount of excess air available was very low. In the case of perfect mixing, this has no effect,
but for incineration, we postulated that the gases leaving the furnace have a heterogeneous
composition consisting of complete separate pockets of oxygen containing combustion
products and "fuel" containing combustion products. At high temperatures "fuel" and air
do not co-exist in a pocket for long, because the combustion reactions occur too rapidly.
When the amount of available oxygen decreases (as temperature increases), the probability
of finding a pocket of "fuel" that has not mixed with available oxygen increases. This
mixing limitation, as indicated by higher CO emissions, was observed for temperatures above
about 1800°F.
Reduction in Emissions
The key to the elimination of combustible emissions is enhancement of combustion
efficiency in the overfire region of the furnace. For a long time designers have recognized
the importance of the three T's - time, temperature, and turbulence. In this report we have
gone one step farther in developing the necessary mathematical models for describing the
various phases of the operation so that the interrelationships between the three T's can be
quantitatively described.
11
-------
t-o
7000
6000
5000
- 4000
@
8
oc 3000
O
2000
1000
t t
Quench
Regime.
1
\
\
\
\
\
.Mixing
Limitation
Zone
X*
1 1 1 1 t* 1 1 ! 1 1 1 1 1 1
700
1000
1500
TEMPERATURE (°F)
2000 2200
Figure 1 Emission factors vs. temperature.
-------
TEMPERATURE CONTROL
The data of Figure 1 clearly show the higher CO emission at temperatures above about
1800°F or below about 1400°F. We believe the high-temperature emissions are caused by a
mixing limitation resulting from a relatively low amount of available oxygen (excess air). As
more excess air is added (consequently lowering gas temperature), the mixing constraint is
eased and the high CO emissions are reduced. In the case of high-temperature emissions, jets
can be used to add a sufficient amount of overfire air to control the excess air within the
given range. In designing such a system, it is important to obtain a uniform "macroscale"
mixing throughout the furnace. The specific design criteria are discussed later in this section.
Using this type of control, the low levels of combustible emissions (as obtained in
Figure 1) can be achieved.
Mixing
As pointed out above, it is important to maintain a uniform macroscale mixing
throughout the furnace, primarily to avoid hot and cold spots, i.e., to maintain the gas
within the 1400 to 1800°F range. However, in terms of burnout of combustible emissions
one must also consider the problem of microscale mixing in addition to the macroscale
considerations of uniformity. The microscale problem is concerned with the mixing within a
local region of the furnace, or along a given streamline, to bring the combustibles and
available oxygen into intimate contact with each other so that the desired burnout can
occur.
To describe this macroscale mixing, it was necessary to introduce two well known
concepts that have not been previously applied to incinerator designs. These are:
(1) The use of the unmixedness factor to characterize the level of emissions; and
(2) The use of statistically based mixing theories to describe the relationship
between the degree of mixing and the time required to burn out a given
pocket of combustibles.
These concepts are discussed in detail in Section V. A qualitative discussion of these
concepts will be included here.
Unmixedness
For complete combustion, mixing on a molecular level is required. A single
sample containing both oxygen and combustibles reflects alternate sampling
from oxygen-rich and fuel-rich gas pockets. A measure of unmixedness of
the overall sample can be deduced from the ratio of the concentration of
oxygen available in the gas sample to that required to complete the combus-
tion process using the statistical description of turbulence. This measure [as
13
-------
suggested by Hottel (4)] reflects both the intensity of turbulence as well as
the relative concentration of the combustibles and air.
The emission data shown in Figure 1 correlate with furnace temperature and
Hottel's (4) unmixedness factor to the extent that the scatter in the data can
be resolved within about 10%.
e Microscale Turbulent Mixing
The final intimate mixing of oxygen and fuel can be described as a turbulent
fluctuation decay process analogous to eddy decay phenomena that have
been extensively studied in wind tunnel experiments.
The mathematical relation suggested by Corrsin (5) equates the unmixedness
factor (^/c'2) to an exponential function of residence time and mixing
intensity (turbulence).
Mixing intensity and residence time are both affected by the furnace design. Since the
unmixedness factor can be related to design parameters, based on Corrsin's equation, and to
the emission factor, based upon Hottel's relation, this defines the overall quantitative
relation between emissions and the three T's. The data taken in this program were
sufficiently accurate to support this conclusion, but not sufficiently accurate to provide
proof of the hypotheses.
DESIGN GUIDELINES
To achieve the desired reduction in combustible emissions, a proper design must
address both the macroscale and microscale mixing effects. The uniform, or macroscale,
mixing can be achieved both through the proper use of overfire jets and through careful
design of the furnace enclosure, considering both the configuration and furnace dimensions.
The jet design strategy is based upon the method of Ivanov (3) which could be easily
verified from the test data. Several tests were run using various jet positions, including
interlaced, opposed, and one-sided jets, and all were shown to be equally effective in
providing uniform macroscopic mixing and temperature control capabilities. The jets proved
to be most advantageous when used to achieve an overall side-to-side mixing and to maintain
the overfire temperature within the proper range.
Overfire jets have been suggested also for use as a means of obtaining front-to-back
mixing in the furnace. However, we believe the same effect would be more easily obtained
through the proper design of the furnace configuration and proper control of the air
distribution so that quenching zones are prevented from occurring, and so that hotter gases
are mixed with the colder gases to attain a uniform temperature throughout the furnace.
The latter is achieved by designing a furnace with flow patterns so the hot gases are
deflected into the colder gases, or in some other way forced to mix with the colder gases,
rather than allowing stratification to occur, as that demonstrated in the testing program.
14
-------
In addition, the furnace should be a sealed enclosure so that maximum control of the
air is maintained through the elimination of leakage. Moreover, the roof of the furnace
should be fairly close to the refuse bed so that the gases rising from the bed do not have a
chance to accelerate under the influence of buoyancy forces. By limiting the velocity of the
gases, the jets will be able to achieve a much greater control of the overall mixing patterns
within the furnace.
Microscale Effects
Overfire jets can also be used to increase the mixing intensity of the overfire gases, but
care must be taken to avoid consequential quenching of the combustion reactions. The
primary use of the jets is for the macroscale mixing and temperature control, but if the jets
can be so designed to accomplish these major objectives and still be capable of adding
sufficient mixing power to the overfire regions to increase the microscale turbulence level,
then so much the better. In our experience during the tests, we found that the amount of
power that would be required to attain a noticeable effect on mixing intensity was
extremely large so as to be prohibitive. The mixing power required to achieve a given mixing
intensity scales as the fifth power of the characteristic dimension of the furnace. Calcula-
tions show that for the large municipal incinerator the effects of jets on microscale mixing
are only marginal, being effective in some cases and ineffective in other cases. But in smaller
furnaces, the power requirements are greatly reduced so that jets could be significant in
affecting the mixing intensities. As a result of the tests, we were able to develop a design
criterion to determine in which cases the jets could be used to attain a mixing intensity
advantage, in addition to accomplishing the macroscale mixing and temperature control
functions. This is discussed in Section VIII. The criterion is based upon the quantitative
relationship between the emission factor and the three T's that was developed during the
course of the program. The full utilization of this criterion and of the other conclusions
derived from this program will represent a major step forward in the development of
incineration technology within the framework of combustion engineering.
15
-------
SECTION V
CONCEPTS IN INCINERATION
Incineration has proven very difficult to describe quantitatively, because of the
simultaneous occurrence of a great number of very complex phenomena. The description of
a combustion process must include fluid mechanics, thermodynamics, heat and mass
transfer, reaction kinetics, and statistics. However, these theoretical principles for the most
part have been reduced to rules-of-thumb when applied to the design of an incinerator. For
example, designers assume burning rates of 60 lb/hr-ft2 and heat release rates of 25,000
Btu's/hr-ft3, basing their designs on past experience or upon attempts to maximize the
effects of the three T's: time, temperature, and turbulence.
In this chapter, we will describe the processes which occur during incineration, using
the mathematical models developed as a part of this and the previous study on overfire
mixing (2). In addition, we will propose a fundamental basis for many of the common
design rules. Required background material related to the basic principles of combustion
chemistry, kinetics, equilibrium, and the like is given in Appendix D.
STRATEGY FOR EMISSION CONTROL
Two phenomena which result in combustible emissions were identified in the course of
this study. The experimentally determined emission factor for the furnace was shown as a
function of furnace temperature in Figure 1. In regions of the furnace where the tempera-
ture was below approximately 1100°F, emissions occurred because the reaction had been
quenched, i.e., it was too cold to sustain combustion.
High levels of emissions are also associated with high furnace temperatures. It is not
uncommon to measure concentrations of CO or H2 on the order of 1000 ppm in flue gases
containing 10% O2 and at 2000°F. These emissions are the result of incomplete mixing, i.e.,
combustibles and oxygen have not come into intimate contact at any time during the entire
passage through the furnace.
The key to the reduction of both sources of combustibles emission lies in proper
control of residence time, furnace temperature, and turbulence. Residence time is a design
parameter of the furnace. During the testing of the existing incinerator this parameter could
not be varied. The dramatic effect of local furnace temperature can be seen in Figure 1. The
third parameter turbulence and the related rate of mixing are more difficult to charac-
terize than either time or temperature. However, mixing is much more controllable in tests
than the other two parameters. For example, the most direct approach to controlling
high-temperature emissions is to add more air; this process is controlled by mixing. As
another example, a possible approach to eliminating cold-temperature emissions is to force
gases emanating from relatively cold areas of the furnace to pass through the hotter gases in
the furnace. This process is also controlled by mixing.
17
-------
EFFECT OF MIXING
In the test program, high-velocity overfire air jets were installed on an existing
incinerator so that the effects of overfire mixing could be more fully explored. During the
course of the work two mixing concepts emerged:
1) Macroscale mixing - required to obtain uniformity of gases throughout the
furnace so as to eliminate or minimize the effects of cold or hot spots; and
2) Microscale mixing required to bring molecules of oxygen and combustible
gases together so that complete burnout of the combustible emissions could
be obtained.
These two scales of mixing are completely different in character and are discussed
separately. However, the optimum strategy for reduction of combustibles is common to
both mixing scales, i.e., supply the available oxygen at selected locations in sufficient excess
to burn out the combustibles in a reasonable time without quenching the reactions.
Macroscale Mixing
Non-uniform mixing throughout the furnace is quite common because of the extreme-
ly diverse combustion conditions observed over individual grates. Proper use of air jets firing
across the fuel bed results in two benefits on the macroscale:
1) The jets supply additional oxygen so that the stoichiometry can be adjusted
to the optimum level for control of high-temperature emissions; and
2) Jets can break up the stratification of the combustion gases so that oxygen-
deficient (hot) regions and oxygen-rich (cold) regions are mixed.
To perform the required mixing functions, a properly designed jet must:
o penetrate far enough into the combustion chamber to influence the macro-
scale flow patterns; and
o carry sufficient oxygen to provide for combustion of the fuel components
present in the gas, but not so much as to quench the reaction mixture.
Careful design is required to satisfy both criteria.
Jet Penetration The design procedure for an overfire jet system was worked out in
the previous mixing study based on the method of Ivanov [3]. In summary, antappropriately
designed jet is one in which the momentum forces predominate over buoyancy forces. Using
these criteria and considering the basic fluid mechanics of turbulent jets, Ivanov presented
an equation for the penetration of the jet as follows:
18
-------
dJuJ
u,.
(3)
where
Lp = penetration distance
p = gas density
dj = jet diameter
u = velocity
and subscripts
j = jet
c = crossflow (in furnace).
Based upon an empirical analysis of jet interactions, Ivanov suggested a jet spacing of
between 3 and 5 jet diameters. Since the flow rate from a given jet can be calculated from
the equation:
«-_^L . .. * LpUc
1.6
(4)
and each jet covers an area of grate:
Ac = (Sdj) Lp
(5)
where
S = the spacing parameter (between 3 and 5),
then the amount of air introduced over a unit area of grate is:
QJ TT »c pr
_ _ J . I ._ f£\
(J. I \OJ
4J Ac 4 1.6S Y PJ V '
which is dependent upon only the crossflow velocity, the spacing parameter, and the density
ratio. The crossflow velocity can be determined from the model for furnace flow as
discussed below.
Furnace Flow In most existing municipal incinerators, flames are concentrated in the
initial portions of the furnace and are considerably diminished toward the end of the grates
where the refuse burns out. Previous models of the fluid-flow patterns in an incinerator have
postulated the existence of stratified zones [2]. Hot furnace gases from the front end of the
19
-------
furnace tend to rise because of buoyancy and exit the furnaces along the roof of the breech.
Colder gases from the after sections form a layer exiting along the bottom of the breech.
The two zones are stable and will not mix. The colder zone is often below the quench
temperature, resulting in high levels of combustibles. Provisions for mixing of the two
stratified zones will create a uniform condition more conducive to combustible burnout.
In our previous mixing study [2], we used a two-zone model to characterize the
differences between the hot gases and cold gases in the furnace. Bernoulli's equation can be
applied to both the hot zone and the cold zone, and the two equations can be combined as
follows:
Hot Zone
Cold Zone
H
PH
gr -2g(z-z0)
(Po - P)
- =(z-zn)
(7)
Combining (APH = APC) yields:
-1
u' = u*
C C0
(8)
The hot gases accelerate in traversing from the bed to the breech as a result of the net
buoyancy force created by the temperature difference between the hot and cold zones. As
the temperature of the hot zone approaches that of the cold zone, the amount of
acceleration is reduced. Note that the cold gases do not accelerate, but leave the furnace at a
constant velocity. If the furnace is not sealed (as in the case of the Newton incinerator),
then there is an interaction between the inside and outside of the furnace such that both the
hot and cold zones accelerate with the hot zone accelerating faster than the cold zone. The
equations for flow in this case are based on the ambient temperature T as follows:
0
UH=UH,
Hot Zone
2g(zH-zJ
Cold Zone
-1
+ 2g(zc-z0)
-1
(9)
Combining the two equations and assuming that (ZH z0) equals (zc z0) yields:
2g
(z-z0)
/TH-TC\"
\ T0 /
(10)
To digress for the moment, note that the time required for the gas to travel a distance
along a streamline can be calculated by integrating the equation about the streamline:
S.
=f dfi(x,z)/uz
(11)
20
-------
The residence time can be calculated from this relation if the path of the gas is known.
An analytical solution to this problem is quite complex and graphical procedures are
therefore preferred. In determining the time required to reach a given position over the bed,
a simplifying assumption that the direction of motion near the bed is always vertical results
in an analytical solution to the above equation for the hot zone:
T =
g(TH - To)
(12)
Note that if I/Hospital's rule is applied in taking the limit as TH approaches T0 , the
equation reduces to:
= (z - zn)/up
(13)
This is a form of the familiar equation:
Time = Distance/Rate
which applies to the cold zone (or any non-accelerating zone) under the assumptions made.
Stoichiometry - The air rate required to achieve the optimum stoichiometry is not at
all a function of fluid mechanics, but rather is dependent upon a difference between the
actual and the optimum concentrations of offgas emanating from the bed. In the section on
microscale mixing, methods will be discussed for estimating these optimum concentrations.
Consider the effect of the jet in mixing air with the offgas from the bed. A material balance
can be made around the zone of jet action as shown in the illustration below. A material
balance, assuming no change in temperature in going from the bed to the jet level, results in
the equation:
where
uc =
u0 =
Ac =
= ucAc
crossflow velocity,
initial velocity leaving the bed,
crossflow area at the jet level, and
bed area.
(14)
21
-------
Hot Zone //Cold Zone'
Qj
Refuse
Since the area of coverage of each jet is fixed by the jet design, according to Eq. (5)
above, then the amount of gas coming from the bed that must be stoichiometrically
adjusted by a given jet can be calculated. Note that if the jets are evenly spaced, then the
jets in the hotter sections of the furnace interact with gases from proportionately larger
grate areas than the jets in the colder section of the furnace because of the "necking down"
process, common with accelerating gas streams. The volumetric flow into the zone of
influence of the jet is therefore equal to uc xA.c . On the basis of ideal gases, a mass balance
can be made about each jet zone as shown in Eq. (15) below:
jiT. QciTc QfiTf
7 + ~, = ~, (15)
where QlT = volumetric flow rate evaluated at temperature T.
and subscripts j, c, and f = jet, cross-flow (before jets), and furnace (after jets), respectively.
Finally, a material balance of the offgas leads to the relation:
Vc -CTc -QclTc= yf -CTf -Qf'if 06)
where y = mol fraction of "offgas," and
CT = molar concentration (CT = P/RT).
and the subscripts c = crossflow condition (before jets), f = furnace condition (after the
jets).
*
This equation can be combined with Eq. (15) and the ideal gas law to derive the relation:
<4IT. / T \
yc- yf _ J_! (A_) m}
O I \ T / (1 '>
Yf g'lTc ^TJ 7
22
-------
By rearranging and substituting for Qc the jet flow rate can be determined by:
This equation differs from the one based on jet penetration criterion Eq. (4), not in the
design variables dj, Lp, and uc, but in the spacing parameter S, the temperature ratio, and
the concentration terms. If the concentration term is determined to achieve the optimum
emission level, i.e., if Cf is the concentration at the minimum point of the U-shaped
emissions curve (Figure 1), then there is no assurance that complete penetration will be
achieved. Likewise, if the jet flow rate is designed for complete penetration, there is no
assurance that the optimum stoichiometry above the bed will be attained. The skill of the
designer is then important in being able to balance the tradeoffs between complete
penetration and optimum stoichiometry.
To further specify the design, note that Eqs. (4) and (18) could be set equal to each
other and simplified to the relation:
6.4S * «' ' 09)
If the composition of offgas emanating from the bed could be predicted, then the left
side of Eq. (19) could be evaluated. The equation would then provide a basis for deter-
mining S or the number of banks of jets that could be used to introduce all of the required
air. Alternatively, the jet penetration criteria could be used to determine the amount of
overfire air required. Equation (19) then forms the basis for determining the underfire air
distribution necessary to produce the optimum stoichiometry. At the present time, our level
of understanding does not allow us this additional degree of freedom.
Microscaie Mixing
The co-existence of combustibles such as CO or H2 and oxygen at high temperatures
suggests that the gas has a heterogeneous consistency, as shown in Figure 2. The rate of
dissipation and eventual burnout of the entrained pockets of combustibles is controlled by
mixing at the microscopic level, i.e., microscale mixing. The problem is to correlate this
mixing mechanism with the reduction of emissions.
The characterization here is adapted from turbulent fluid mechanics and mixing
studies in high-velocity burners. Two distinct concepts are required to complete the
characterization:
1) The relation of emission factors to fluid mechanics concepts, and
23
-------
2) The relation of fluid mechanics concepts to incinerator design parameters
(the three T's).
7
COMBUSTIBLE POCKETS
AIR POCKETS
Figure 2 Heterogeneous character of flue gases
Emission Factors and Fluid Mechanics The mixture of complete and partial combus-
tion products emanating from the refuse bed is a gaseous "fuel," and the mixing process can
be characterized as being similar to the mixing in turbulent flames (as suggested by
Hawthorne, Wendell, and Hottel [4]). The characterization which results in a statistical
description of turbulence can be summarized as follows.
In turbulent systems, concentration at any point in the system can be represented by
two components: (1) the time mean value concentration, and (2) the fluctuating component
of the concentration. These are represented by the equation:
C = C + c' (20)
where c' has been shown to be randomly distributed about C. The units of C can be either
moles/volume or mole fractions.
By applying the gaussion distribution to c', Hawthorne et al (4) showed that the ratio
$, of oxygen in a sample to the oxygen required to complete the combustion of the sample
could be used to determine the values of the parameter y/c'2/(CCs)* where-y/c^is the
r.m.s. value of c' and (C-CS) is the difference between the initial "fuel" concentration C
before any combustion occurs and the stoichiometric concentration Cs,i.e., that initial
concentration of "fuel" which would just consume all oxygen present. Note that the
variable C represents the concentration of "offgas" in air and not the concentration of the
'Hawthorne's derivation uses C to represent either mol fraction or molar concentration interchangeably.
The same convention is used throughout this report where C is often used interchangeably with y, the mol
fraction of offgas.
24
-------
constituents making up the offgas. Also, note that C is not an average value of the
concentration like Cs, but rather an instantaneous value which is a function of time and
position.
Hawthorne et al (4) further showed that which can be calculated from the sample
analysis, can be used to determine the nature of the unmixedness factor^/c2^ based upon
the relation shown in Figure 3. The derivation of this relation is in Appendix F. An example
of this calculation is given in Appendix B. The relation will be shown later to correlate quite
well with the emission data for incinerators.
s
0 40 *0
MO MO 400
ACTUAL OXYGEN IN SAMPLE
02 REQUIRED TO COMPLETE COMBUSTION
Source: Ref. (4).
Figure 3 The unmixedness factor y c'2 in turbulent flames
The value (AC) can be calculated if the original offgas composition is known. For pure
fuels, the calculation is straightforward; for incinerators, the composition of offgas ema-
nating from the refuse bed must be assumed. This is examined in a later section of this
report. However, if AC_can be determined, then Figure 3 can be used to determine the
unmixedness factor, VA^27 Hottel (4) has proposed that this factor be used as the character-
ization factor for c' to relate emissions to fluid mechanics.
To illustrate the relation of this concept to emissions, consider the time distribution of
concentration, shown schematically in Figure 4, where C is the concentration of refuse bed
offgases in air before combustion, Cs is the stoichiometric concentration of offgas in air, and
25
-------
Cq is the concentration* of the offgas at the onset of quenching. Concentrations which
exceed Cs represent regions which, if well mixed, would not have sufficient oxygen for
complete combustion. Any such regions surviving until the gases leave the furnace will
contribute to the total emission of combustibles. Likewise, concentrations below Cq
represent regions where combustion has been quenched as a result of low temperatures. If
these regions contain combustibles, then they too contribute to emissions.
u.
H UJ
<0 C
Zc/5
UJ _1
O UJ
I2
MEAN
HIGH-TEMPERATURE EMISSION REGION
LOW-TEMPERATURE EMISSION REGION
TIME
'Expressed as fraction reduced to unreacted constituents.
Figure 4 Concentration fluctuations.
The source of emissions and the unmixedness factor come together when one notes
that the unmixedness factor, v c'2 . is the standard deviation of concentration fluctuations
about the mean concentration. If the distribution is gaussion, as Hawthorne et al (4) suggest,
then C and vc71 are sufficient to determine the entire distribution. Furthermore, one can
determine Cq and Cs from the composition of the offgas, so that the amount of gases within
the concentration distribution falling outside the region from Cq to Cs can be estimated.
The techniques for control then become much better defined.
From the analysis of the two emission-producing conditions discussed above, it is clear
that combustible emissions can be minimized or avoided if the concentration fluctuations
*Since concentration and temperature are directly related, a quench temperature can be represented by a
corresponding concentration.
26
-------
are maintained entirely between Cs and Cq. This can be done either by adjusting the local
air-to-fuel ratio or by increasing the level of turbulence. These two approaches, illustrated
below, in Figures 5 and 6, are the province of microscale mixing.
With respect to the stoichiometry, consider the distribution of fluctuations in fuel
concentrations shown in Figure 5. In distribution A, the mean concentration is close enough
to Cq so that a large portion of the distribution lies within the quench zone and emissions
resulting from low furnace temperatures are expected. In distribution C, on the other hand,
the mean concentration is close enough to Cs so that a large portion of the distribution has a
concentration greater than Cs and emissions will result because of the mixing limitation.
Clearly, the emissions can be minimized, as in the case of distribution B, by adjusting the
mean concentrations so that the portion of the distribution outside of the range from Cq to
is minimized.
"s
The above example illustrates that the value for C can be shifted between Cs and Cq
without changing the level of turbulence in the system simply by adjusting the excess air in
the furnace, i.e., by controlling the furnace temperature. Of the two variables used for
controlling emissions, this one is by far the most useful.
The effect of the level of turbulence is shown in Figure 6. The greater the level of
turbulence, the narrower the frequency distribution becomes, increasing the probability of
operating the furnace with no concentration fluctuations outside the combustion zone
range. In distribution A, the mixing is poor; it is increasingly better in distributions B and C,
respectively. The limiting case, of course, is that of infinite mixing (distribution D), resulting
in zero emissions for all concentrations between Cq and Cs.
Fluid Mechanics and the Three T's A considerable amount of research has been
devoted to the study of turbulent mixing, primarily based on the statistical theory of
turbulence as applied to the aeronautics industry. Correlations between velocity fluctuations
behind a grid in a wind tunnel and stream parameters have been thoroughly studied. As a
result, the decay of eddies (represented by eddy velocities) with respect to time is both
empirically and theoretically described. This work is reviewed in References 5 through 8.
Relying upon the familiar similarities among heat and mass and momentum transfer,
Corrsin (5) suggested the concentration analogy to velocity decay, reasoning that fluctua-
tions in concentration and velocity both follow the same equations. The fundamental decay
equation is:
yc^X/cf" = e-# . (21)
The factor 0 is defined by the equation:
(22)
27
-------
MEAN
MIXING
LIMITATION
ZONE
Q:
Figure 5 Effect of stoichiometry on emissions for a fixed turbulence level.
MIXING
LIMITATION
ZONE
Figure 6 Effect of turbulence level on emissions for a fixed stoichiometry.
28
-------
where
L = the characteristic dimension of the system,
e = the rate of input of mixing energy per mass of system, and
k = 0.19, as determined from turbulence experiments in wind tunnels [5].
Since the term y/c'2 can be related directly to the emission factor of the furnace
(through the stack gas analysis), this equation represents the relation between emissions and
both time t and turbulence as characterized by (3.
The effect that micromixing of this nature has on emissions is shown schematically in
Figure 7. In the first diagram, the fluctuation in fuel concentration with time is shown.
Because concentration fluctuations will eventually decay (according to the Corrsin model),
then as long as the mean concentration C falls between Cs and Cq, the entire distribution of
fluctuations will eventually fall into the combustion zone and the emissions will be
eliminated. To promote this change more rapidly, one can follow two different courses of
action:
Shift C relation to the constraining boundaries Cs and Cq, or
Increase 0t to achieve a faster rate of decay in the fluctuations.
The first approach to eliminating emissions on the microscale is shown in the second
diagram of Figure 7. Whenever the mean concentration is close to Cs (or alternatively to
Cq), then the emissions will tend to be high. However, by adjusting the mean concentration,
a greater proportion of the fluctuations will occur within the combustion zone and the
emissions will be diminished. Since the concentration is directly related to the temperature,
the importance of maintaining the temperature within the combustion range is clearly seen.
This is the primary effect of temperature on emissions. The second approach focuses on
achieving more complete decay of the concentration fluctuations.
In the third diagram, a typical concentration fluctuation for a relatively unmixed
system would appear as the solid lines and would result in a given amount of emissions at a
specified concentration (or temperature) and |3t. By maintaining the mixture at the same
concentration, but improving the level of mixing, the fluctuations can be reduced as shown
by the dotted profile so that the emissions can be reduced or eliminated. The same
argument also applies to increases in residence time at constant temperature and degree of
turbulence. Although the singular importance of adjusting stoichiometry is crucial to the
reduction of combustible emissions, these secondary relationships between temperature and
both turbulence and time have also been noted in order to complete the quantitative
description of the interrelationships among the three T's.
Application of Micromixing Models to Incineration The previous two headings of
this section dealt with the development of a micromixing model to correlate emissions data
29
-------
OJ
o
^STOICHIOMETRIC
.EMISSIONS
MEAN
Q
ENCH
EMISSIONS
NATURAL DECAY
ADJUSTMENT OF STOICHIOMETRY
PROMOTION OF DECAY
Figure 7 Overall effect of micromixing on emissions.
-------
with furnace operating parameters, particularly the three T's. The two parts of the model
that were presented are:
Use of the unmixedness factor, >/c'2 to relate emissions data to fluid
mechanics properties of the system; and
Use of Corrsin's Equation to relate the fluid mechanics properties to time,
temperature, and turbulence.
In this section, we have outlined the steps required to apply both parts of the model to
incinerators.
Throughout this part of the report it is necessary to understand the combustion
phenomena at work within the incinerator. In the previous mixing study (2), we developed a
model to characterize the bed burning process. During the course of this work, it became
more and more apparent that the characterization of bed combustion was not one of the
primary factors necessary to understand and control combustible emissions. However, the
bed model was often necessary to estimate the value of the parameters used in the mixing
models. The development of the bed burning model is given in Appendix E. Some of the
results of this development are summarized below.
Within the refuse bed, temperatures are high enough to assure complete burning of
pyrolysis or gasification products that come into intimate contact with oxygen. More likely,
however, there will be insufficient oxygen or poor mixing so that combustible gases will be
given off.
The offgas contains two distinct constituents: (1) combustibles and combustion prod-
ucts, and (2) air. The composition of the combustibles and combustion products is
controlled by the water gas shift equilibria. The composition can be estimated (as shown in
Appendix E) if the initial value of CO/CO2 is known. Typical compositions are shown in
Table 1 for different values of CO/CO2. These are also shown graphically in Figure 8.
The actual value of CO/CO2 was determined empirically to be 0.21 (range 0.02 to 0.6),
based on test data taken directly over the bed of an existing incinerator. The data are
presented in Section VII.
The offgas from the refuse bed also contains oxygen that has not yet been consumed.
Since oxygen and combustibles do not co-exist in a well mixed state, the oxygen available in
the offgas must be contained within a pocket of unmixed air. The fraction of such air in the
offgas is simply the ratio of available oxygen in the sample (in percent) to 21%. In the tests
reported in Section VII, the incinerator contained between 29% and 90% unmixed air with
an average of about 50%.
31
-------
Table 1. COMPOSITION OF BED OFFGAS
CO/CO,
0.25
0.50
1.0
1.5
2.0
Analysis (vol %)
C02
CO
H2O
H2
N2
13.5
3.4
22.4
2.8
57.9
12.2
6.1
22.0
5.5
54.2
10.5
10.5
21.0
10.5
47.5
9.3
14.0
19.9
15.0
41.8
8.4
16.8
18.9
18.9
37.0
02 (req.) 3.1
mol 02/100 mol offgas
5.8
10.5
14.5
17.8
30
20
LLJ
O
oc
111
Q.
10
0.5
I
1.0
[CO/C02]
1.5
2.0
Figure 8 Offgas composition of CO/CC-2
32
-------
Application of the Unmixedness Factor
In Appendix B, the procedure for calculating the unmixedness factor from a stack gas
analysis is given. It is also possible, given an unmixedness factor, to calculate the emission
factor if the initial CO/CO2 ratio is known. In Figures 9, 10, and 11 are shown three sets of
calculations based upon different CO/CO2 values and also different empirical relations for
CO2 vs temperature. In this work both the initial CO/C02 ratio and_the stack gas analysis
were measured. Hence it was possible to determine the value for >/c'2 two different ways
and compare the resu'ts. This is also done in Section VII. The good agreement between the
experimental value of >/c'2 (based on gas analysis) and the theoretical value of\/c'2 (based
upon the CO/CO2 ratio) supports the contention that emissions can be determined using the
unmixedness factor approach.
Application of Corrsin's Equation
To apply Eq. (21) above to incinerators, one must be able to estimate the following
parameters:
The first two are related to the aerodynamics of the furnace mixing and also to the furnace
residence time (for gases, not refuse). The latter two parameters require the knowledge of
the composition of offgas emanating from the refuse bed. The estimation of each of these is
discussed below.
Parameter 0 For the incinerator, the only types of mixing energy that have been
considered are those introduced by incoming air streams. Equation (22) can then be
expanded as follows. The rate of input of mixing energy for a given air stream is calculated
from the equation:
d (KE) u mil2 ~*, , <
u_£_; = AJ321L = '/2(pAu)u2 = V2pQ3 A2 (23)
dt At
If each stream is evaluated at the temperature at which.it gives up its energy to create
turbulence, then as a conservative estimate each jet must be evaluated at the jet tempera-
ture, while the gases coming off the refuse bed are evaluated at the furnace temperature.
The energy term for a jet acting over a portion of the grate is:
e = p
^ :
33
-------
2500
CM
8
o
o
Q.
tr
O
o
to
CO
2000 -
1500
1000 -
500 -
750
1000
1750
1250 1500
TEMPERATURE, °F
Figure 9 Dependence of emission factor on temperature and unmixedness factor.
2000
34
-------
3000"
2500 -
CM
O
CJ
S9
CM
@
O
O
a
a
ir
O
u
2000 -
1500 -
1000 -
BASIS:
C0/C02 = 0.5
%C02 (WET) =
-------
2000
1500 ~
CM
8
8
a
cc.
O
z
O
1000 -
500 ~
BASIS:
C0/C02 = 0.50
%co2|w
= (TEMP-750UF)/1250UF
1000
1250 1500
TEMPERATURE, °F
1750
2000
Figure 11 Dependence of emission factor on temperature and unmixedness factor.
36
-------
By substituting in the ideal gas law and scaling all dimensions with the characteristic
dimension L, which is taken to be the shortest dimension of the furnace (usually the furnace
width), the equation for j3 becomes:
QJ
Qo
1/3
(25)
where
7 :
subscripts
Lf/L = ratio of length to width,
Hf/L = ratio of height to width, and
o = initial condition at top of bed
j = Jet.
If uniform mixing has not been achieved, then j3 will have a different value for each furnace
streamline.
Parameter t Parameter t is the residence time of combustion gases within the
furnace. One method for calculating t was proposed in the section on fluid flow [see Eqs.
(11) through (13)]. Another approach is to relate t to the gas flow rate Qf and the furnace
volume Vf. This could be combined with Eq. (25) above to calculate j3t according to the
equation:
= 0.15
assuming that
(26)
t =
Note that without jets (0t)0 equals 0. 1 5 y2 / 3 .
In this case, the dependence on furnace length is eliminated. Since macroscale uniformity is
more probable across the width of the furnace or from grate to roof, but not from back to
front, the use of this form of the equation is more consistent with the local character of the
microscale model.
Parameters
and AC The effect of overfire air jets can be determined from
a material balance of the jet mixing process using the calculated offgas composition for
37
-------
refuse beds (See Appendix E). Consider for example how the composition of bed offgas is
affected by dilution with high-velocity air jets. The following equations are based upon the
point of jet mixing as sketched below:
Jets
..;;..- Diluted-.;
"'.'.''.. Gases ".
*>-
Jet Mixing Zone
.."..Offgas .-.
Refuse Bed
Let x equal the mole fraction of offgas in the diluted gases. Then x can be estimated in one
of two ways. The amount of carbon in the offgas is the same as the amount of carbon in the
diluted gases. At the breech of the incinerator essentially all of the carbon is in the form of
CO2. If a gas analysis is taken before and after the jets are turned on, and if subscript b
represents the base condition (before jets) and subscript j represents the jet condition, then
the carbon concentration represented by CO2 is:
lj = x(%C02lb)
or
x =
%C02lj
%co
(27)
(28)
2'b
Alternatively, a balance of total moles at the point of mixing yields the equation:
PQi PQC PQf
RT,
RTf
(29)
In this equation, the variables that can be measured are Qj, Tj, Qf, and Tf. The defining
equation for x then becomes:
x = 1 -
Qf/Tf
(30)
The factors required to calculate yc'2 all change with the addition of jet air. The
changes in these variables are calculated using the following equation. The change of )3t is
best determined based upon Eq. (30):
1 +
(Af/Tf)
0-x)3
1/3
(31)
38
-------
Addition of air will increase available oxygen according to the equation:
(%02 )j = x (%02 )b + (1-x) 21%
(32)
Likewise the oxygen required for complete combustion after dilution is determined from
the equation:
(O2Req)Jo =(02 Req)bo-x
(33)
Dividing Eq. (32) by Eq. (33) and combining terms, one obtains the following:
l-x\ 21%
/ %0a\ / %02 \
\°2Req/J0 \02Req/b
1 +
(%02)b
(34)
Note that Eq. (34) applies only to the region of jet mixing (subscript o), not to the entire
streamline. A similar approach is taken to determine Ay. In Appendix B it is shown that:
%CO2
(35)
%C02lb\
%co2ij
where y represents the mole fraction of original combustibles (before air entrainment within
the bed). The addition of jet air carries the dilution of the offgases even further than occurs
within the bed itself. At the exit of the bed:
%C02b
(36)
After the jets:
(y-ys)j = ys
/%CO2I
\%CO2
The equation for Ayj can be expressed in terms of x as:
'%C02 b
%C02
(1-x)
(37)
(38)
The above equations can be combined with Corrsin's Equation and Figure 3 to provide a
single model equation as follows:
From Figure 3:
'2
Ay
(39)
39
-------
Note that f (00) is given in Appendix F.
Combining this with Eqs. (34) and (38) yields:
The jet effect can be graphically illustrated in the following way. Figure 10 can be
replotted (Figure 12) to show emission factor vs \/c'2 (rather than temperature). Then
Point A on Figure 12 represents some base condition as determined in the breeching of the
furnace. The initial base condition can be determined by applying Corrsin's equation in the
form:
The initial point can be located on Figure 12 by assuming that the initial and breech
conditions are at the same temperature, i.e., the decay process is isothermal as shown in
Figure 1 2 (Point B). One could also apply the model by specifying the initial temperature
rather than making the isothermal assumption. The initial jet condition can be determined
by noting that the emission factor does not change with dilution, but remains constant in
going from the base to the jet condition. Because of the isothermal assumption, the factor
(v/c2 Y is the value determined at the emission factor of the base condition and the
V /Jo / r=.\
temperature of the jet condition, i.e., Point C. The decay of(-^/c )j is also isothermal, so
that the final value will lay somewhere along the jet isotherm, Point D, for example. The
magnitude of decay can be determined from the equation:
(42)
This equation is the basis for predicting the effect of jets on combustible emissions.
Estimated Effect of Jets on Incinerator Emissions
The effect that jet mixing has on combustible emissions follows from Eq. (40) above as
seen in the following example.
Consider a furnace initially at 1800°F, having an unmixedness factor of 0.20 and with
j3t equal to 0.25. The relationship between emissions, temperature, and y^c5 is assumed to
be that of Figure 9. The empirically derived relationship between temperature and %CO2 is:
%CO2 x(%CO2 b) T-960°R
= =
10% 10% 1500°R
40
-------
3000
2500
2000
CC
O
O
1500
1000
500
INITIAL VALUE
WITH JETS
EFFECTOR JETS
(SEE TABLE 2)
FINAL VALUE
WITH JETS
.05
.10
.15
.20
.25
.30
.35
.40
Figure 12 Dependence of emission factor on temperature and unmixedness factor (revised plot of Figure 10)
-------
At 1800°F (2260°R), %CO2 is 8.7%. The ratio of Af|Aj is taken to be 150. A summary of
the calculated terms for the jet conditions is given in Table 2 based upon the following
calculations:
Parameter Source
Given
Eq. (27)
Eq. (43)
Figure 3
Eq.(31)
Eq. (40)orEq. (41)
L/d* L - U/c2 } By substitution of Eq. (42)
Note that the internal consistency of the equations can be demonstrated by calculating the
final two parameters listed above, using either of two equations each.
The values for ye'2 vs temperature given in Table 2 have been plotted on Figure 12.
The jets have a dramatic impact on reduction of emissions whenever x is less than 0.95. In
Section VII, we have shown that other conditions could be set forth in which the jets prove
to be completely ineffective or of marginal value.
42
-------
Table 2. SUMMARY OF CALCULATED TERMS
X
Base
Jet
X
1.0
0.98
0.95
0.90
0.85
0.80
0.75
T,°F
1800
1778
1740
1675
1609
1544
1479
*o
27
28
30
34
38
43
48
/ /Tf\
(v£E)
\ AC/o
0.774
0.765
0.750
0.725
0.700
0.680
0.665
AYO
0.332
0.341
0.355
0.377
0.400
0.422
0.445
/ \
/ /=j?\
r>
0.257
0.261
0.266
0.273
0.280
0.287
0.296
C"~ \ ,
150 .T:\2
1 1 ,
Tf ) <1-*> J
1.0
1.004
1.055
1.356
1.853
2.442
3.032
1/3 ^_
/&
Vc
0.200
0.203
0.204
0.195
0.176
0.156
0.137
A»/c^ *
t* V
0.057
0.058
0.062
0.078
0.104
0.131
0.159
-------
SECTION VI
EXPERIMENTAL PROGRAM
A large portion of this program was devoted to testing an existing municipal inciner-
ator to:
1. verify the validity of applying the basic combustion fundamentals presented
in the previous chapters to the design of municipal incinerators;
2. demonstrate the effectiveness of overfire jets in reducing combustible emis-
sions; and
3. provide experimental data on municipal incinerator operation in hopes of
promoting further technological advances in design or operation.
The test equipment procedures and data reduction methods are discussed in detail below,
with the data tabulated in Appendix A.
TEST INCINERATOR
The primary piece of equipment used in the test was a municipal incinerator. Before
selecting the test vehicle, we reviewed the various types of incinerators in use in the United
States today. We limited our selection to grate-type incinerators, because the batch-type
incinerators are being phased out around most of the country and suspension-fired units are
not in sufficient use to have general applicability. We considered several local incinerators
before finally selecting the incinerator of the City of Newton, Massachusetts.
The Newton incinerator is a 500-TPD capacity plant consisting of two identical
250-TPD traveling grate furnaces as shown in Figure 13. The incinerator configuration is
typical of many found throughout the country today. The refuse arrives at the facility and
is weighed before being dumped into the loading pit. A fine water mist is periodically
sprayed over the refuse to control the dust. Unfortunately, much of this water settles to the
bottom of the pit so that the refuse there is very moist and burns poorly. Both furnaces are
fed by a 3.5-ton traveling bridge crane provided with a 2-yard clamshell bucket. The charge
chute to each furnace is water-cooled and gravity-fed. During operation, the crane operator
is reasonably careful in blending the various types of refuse to obtain a uniform mixture.
Even so, large fluctuations in the refuse composition are seen so that consistent operation
on a particular type of refuse is difficult to maintain for more than one or two hours. In
collecting data on emissions, we preferred test runs lasting for as long as 4 hours, but in
many cases changes in refuse composition or other adverse conditions forced a change in the
operating conditions of the incinerator.
45
-------
ROOF £L. 92.0'
DUMPING AMD STOKII6
FLOOD EL. 63.0'
GRADE EL. 60.0*
O\
"H
Figure 13 Cross section of Newton 500-TPD Incinerator showing sampling locations.
-------
Furnace
Each of the rectangular furnaces is lined with conventional refractory, except for the
lower portion of the wall about three feet above each grate which is lined with a silicon
carbide brick to resist abrasion and slagging. The distance from the grate to the roof at the
charging end is approximately 10 feet; at the drop-off end it is approximately 30 feet.
The furnace has three traveling grates to convey the refuse through the incinerator. The
initial grate is approximately 11 feet long and is inclined to facilitate radiative transfer of
heat to dry the incoming refuse. Hence, it is designated the drying grate. The refuse is drawn
from the feed chute into the furnace through an opening 4.2 feet high by 8 feet wide. The
feed rate of refuse into the furnace can, therefore, be determined by the speed of this first
grate and the density of the refuse. For an average density of about 15 lb/ft3, a grate speed
of 40 ft/hr results in a feed rate of 250 tons/24 hour day, which is the design capacity. The
grate itself consists of rows of overlapping keys through which air is passed both to promote
burning and to cool the burning surface. The keys are made of cast iron to resist high
temperatures. The rack bars holding the keys are made of mild steel. Throughout the
program conditions often existed that, if not properly adjusted, resulted in overheating of
the steel portions of the grate. Since a severe overheating could have caused damage to the
grates, these conditions were abandoned whenever the grate temperature began to rise
abnormally.
At the end of the drying grate, the refuse tumbles on to the second or burning grate. The
vertical distance between the two grates is approximately 5 feet which is sufficient to promote
a certain amount of refuse mixing, which tends to expose additional unburned refuse to the
flame. The second grate is 16 feet long and slightly inclined. Its speed can also be
controlled as can the speed of the third grate. In general, the second grate is operated at a
somewhat faster speed than the drying grate because past experience has shown that when
the grate speed is equal or slower than the drying grate, the latter tends to overheat. The
most probable explanation for this phenomenon is that the slower grate speed of the lower
grate results in a buildup of refuse directly under the nose of the upper (drying) grate.
Then, either because of a restriction in the air leaking into the furnace which normally cools
the nose of the grate, or because of increased heat generation at that point on the grate, the
grate overheats. Since this portion of the grate is not cooled by underfire air, the grate will
return to the charge chute hotter than normal. When this occurs, the refuse on the grate
eventually reaches a point at which it could be prematurely ignited. With the flame right
next to the grate surface, an abnormally high heating effect occurs which is not relieved by
the underfire air. This phenomenon happened several times during the course of the tests
and prohibited operation for long periods of time at relatively slow burning grate speeds.
47
-------
The refuse is tumbled for a second time from the burning grate to the third or
burnout grate. The flame on this grate is characteristically lower in intensity than
on the second grate and, in many cases, the flame has either been extinguished or is
merely smoldering by the time the refuse has traveled the 16-foot length. One characteristic
of this phase of the operation is that many of the slower burning items, such as heavy books
or logs, are still burning when they reach the end of the grate. To provide a longer residence
time for the furnace, this grate was typically operated at speeds equal to or slower than the
second grate. Unfortunately, at slower grate speeds the residue tended to build up,
restricting the air flow through the bed and, consequently, reducing the combustion inten-
sity. A consistent problem throughout the test was the development of methods for
operating this grate to achieve good burnouts, while at the same time maintaining temper-
atures above it that would be well above the quench temperature. The latter is extremely
difficult and consequently temperatures over this grate as low as 800°F were not
uncommon.
When the refuse reaches the end of the burnout grate, its ashes fall into a water-filled
hopper where they are quenched. The residue is next transferred to a truck by means of a
continuous-chain flight conveyor, and eventually is used as landfill. The overall volume
reduction is typically between 80% and 90%.
Combustion Air System
The primary source of combustion air is a forced draft fan rated at 40,000 cfm (6
inches SP) and fitted with an inlet vane control. The underfire air is supplied to. the grates
-by three separate ducts running from the main fan discharge to the air plenum located
between the upper (forward) and lower (return) sections of each of the grates. Each plenum
is divided into two sections, with the distribution controlled by a manual damper at the
entrance to each system. Air is forced through the upper portion of the grate, passing
through the keys and cooling the burning surface. The boxes are not particularly well sealed,
so that high static pressures in the wind boxes cannot be attained. For this reason the
amount of air that can be forced through the refuse bed is severely constrained, particularly
in the drying grate. The third grate has very low refuse bed depths at normal or high speeds
and can accept relatively large portions of air without difficulty. This configuration often
results in high excess air over this grate and consequently low temperatures. ;
The overfire air is taken from the fan discharge using only one large duct with a single
manual damper. The overfire air is introduced into the furnace through three banks of roof
jets consisting of five 6-inch diameter jets each. The primary purpose of this overfire air
system is to cool the overfire region. The overfire mixing system, installed for emission re-
ductions, had a greater number of jets operating at higher jet velocities.
Additional air is used to cool the silicon carbide portion of the wall to prevent slagging.
The air enters the furnace at approximately 3 feet above the grate level. The forced draft fan
supplying this air is rated at 4500 cfm (6 inches SP). This source amounts to less than 10%
of the total air flux into the furnace.
48
-------
By far the largest source of air is in the form of uncontrolled leakage into the furnace.
The leakage rate was calculated by material balance and, typically, ran as high as 30,000 cfm.
A large fraction of this air is drawn in from beneath the grates and is a natural source of
cooling for both the grate and the overfire region.
Stack Gas Removal
Each furnace is equipped with a fly-ash removal chamber containing water sprays and
fire brick baffles. The hot combustion gases pass through the breeching area of the
incinerator into these chambers. The gases are cooled to about 350°F and the large fly ash is
removed. Water from the baffles is collected in a sump, clarified, and recirculated. The
baffle chambers merge directly into a single natural draft chimney which is 175 feet high
and designed for the gas flow required to accommodate the ultimate capacity of both
furnaces. Because the cooling of the baffle system maintains the exit temperatures in a fairly
narrow temperature range of approximately 250° to 450°, the chimney draft that results is
relatively constant. For this reason, both the draft pressure in the furnace and the gas flow
rate through the furnace remain constant. Hence, the air distribution could be somewhat
controlled, but not the flow rate. This constraint was a constant source of operational
difficulty throughout the tests.
Incinerator Control
To control the incinerator at the proper conditions, additional instrumentation was
required to augment its present monitoring system. The grate speeds were determined with
the existing rate meters, and the speeds were verified to ensure that the meters were reading
properly. Pitot tubes were installed in each of the underfire air ducts and in the main
overfire duct to record the air input into the furnace. The total air volume was checked with
a pitot tube reading at the inlet duct to the main fan. In addition, the leakage air was
estimated using a hot-wire anemometer to determine the air velocities at the most probable
points of entry.
INCINERATOR MODIFICATIONS
To follow the strategy for eliminating combustible emissions on both the microscale
and the macroscale, one of the furnaces was modified with an overfire jet system. Based on
information gathered during the initial test of this program, it was clear that the gases from
grates 1 and 2 were very hot and oxygen-deficient, while the gases coming from grate 3 were
at or below the quench temperature. To demonstrate the design principles and effectiveness
for jet systems, we installed sidewall jets over grate 2 to satisfy the stoichiometry suf-
ficiently and also to achieve sufficient penetration. The forced-draft fan that was selected
had a design capability of 15,000 cfm at 6 inches SP. To achieve the greatest flexibility and
control, the fan was equipped with both inlet vane control and an outlet damper.
49
-------
The duct network consisted of 16-gage welded tubing and channeled air from the
outside through the fan to a Y-connection, where the air was split into two streams; each
stream was delivered to a jet manifold located on opposite sides of the furnace. The air
distribution was monitored using a pitot tube in the main inlet duct and pitot tubes in each
of the two ducts feeding the jet manifolds. In addition to the controls on the fan, a
butterfly damper was installed in each of the two manifold ducts.
The manifold system consisted of a long header off which 10 duct sleeves ran, carrying
air to the 10 jets on each side of the furnace. Each tubular sleeve had a 5-inch diameter and
was equipped with both a 1/8-inch velocity pitot tube and a butterfly damper for velocity
control and monitoring. Because of the experimental nature of the program, the duct sleeves
were connected to the jets by flex-hose connections. During the course of the tests, the flex
tubing caused many maintenance problems. Thus, this type of construction is not recom-
mended for permanent installations.
The jet nozzles were constructed of 5-inch, schedule 40, mild steel pipe, inserted
through the wall of the furnace to within 2 inches of the inner refractory wall. The external
portion of the pipe was connected to the duct sleeves by means of a flanged connection. In
installing the jets, the location for each jet was marked on the inside of the furnace, and the
hole was made using standard techniques. The jet pipe was inserted through the wall, welded
to the outer furnace shell, and surrounded with a high-temperature insulation. The inner
refractory wall was then molded around the jet using a plastic refractory. During the course
of the tests, slag buildup around the jets was quite common. Therefore, a permanent jet
installation should use precut silicon carbide refractory or other slag-resistant material to
avoid this problem.
CONDUCT OF TESTS
The composition of refuse that is burned in the average municipal incinerator varies
from hour to hour and day to day, making it difficult to arrive at any consistently
reproducible test conditions. During the operation of the Newton incinerator, these varia-
tions could be seen in breeching temperatures fluctuating as much as 500°F within a period
of 30 minutes; variation in the breeching temperature of ± 100°F was considered a relatively
stable condition. It was therefore difficult to measure precisely all of the variables affecting
the system.
Test Parameters
Test variables included the refuse feed rate, moisture content, the underfire and over-
fire flow rate and distribution, and the refuse residence time. Unfortunately, the test
variables are interrelated and are often dictated by the permissible or possible conditions
under which the incinerator can be operated. Each of these parameters is discussed in de-
tail below.
50
-------
Feed Rate
The refuse feed rate is controlled by the grate speed of the drying grate and it typically
runs between 35 ft/hr and 55 ft/hr, being set according to whether the refuse is relatively
wet or dry. The volumetric rate in cubic ft/hr is a product of the grate speed in ft/hr and the
area of the charging port (33.6 ft2).
The weight of refuse depends upon the refuse density which often varies. The density
of Newton's refuse was reported previously by Kaiser's[8] to be 13 lb/ft3, but similar data,
both at Newton and for other incinerators, indicate densities as high as 17 lb/ft3. In the
calculation of feed rates there is an additional uncertainty due to the packing of the feed
hopper; if it is packed loosely with large voids, the apparent bulk density will be lower than
13 lb/ft3. On the other hand, a tight packing will result in a higher apparent bulk density.
The density also varies with moisture content, tending toward the higher values for wet
refuse.
In the conditions which were used in the current work, the following feed rates were
utilized:
Grate speed, [ft/hr]
Volume feed rate, [ft3/hr]
Mass feed rate,* [tons/hr]
Heat release,** [MM Btu's/hr]
35
1176
7.7
68-78
40
1344
8.7
78-87
45
1512
9.8
88-98
50 55
1680 1848
10.8 11.9
98-107 108-118
*P= 13 lb/ft3
**heating value of refuse = 4500-5000 Btu/lb, 100% burnout.
Moisture Content
Refuse will always contain moisture but it can at times have considerably more
moisture than usual. Typical values range between 20% and 50%. At Newton, the higher
moisture conditions result usually after a rain, or when the refuse is taken from the bottom
of the pit. To control the dust, the pit is periodically sprayed with water while trucks
unload. Excessive moisture drains to the bottom of the pit. At times water was observed
dripping from the clamshell bucket as refuse from the bottom was charged. The effect of
moisture was clearly shown during the runs with the amount of refuse moisture changing
over the course of a day's run as the crane worked from the top to the bottom of the pit.
51
-------
In describing refuse, we have qualitatively used the labels of "wet" or "dry," but we
recognize that the absolute moisture content is not such a discrete function. We also
recognize that our visual observations could be quite misleading, and that the data on the
combustion products could contradict our initial moisture labels. Of the principal variables,
this one is by far the most uncertain and least controllable of the operating variables.
The moisture content was allowed to vary normally, resulting in conditions for the first
tests fairly evenly divided between "dry" and "wet" refuse. Some of the runs reflect a
change from dry to wet during the progress of a test from beginning to end. The final series
of tests demonstrating the use of overfire jets had primarily dry refuse.
Unfortunately, the refuse moisture content, which drastically affects the refuse heat
value, also affects very much the allowable feed rate. Wet refuse is normally burned at a
slow refuse feed rate to achieve an acceptable burnout. If higher feed rates are attempted,
the ash level on the burnout grate tends to build up, and the ash conveyor system becomes
overburdened. Continuous operation in this mode during some of the tests resulted in
temporary shutdowns because of the strain on the ash conveyor system. On the other hand,
by reducing the second and third grate speeds a good burnout could be attained, but then the
furnace would overheat. Hence, wet refuse was almost always run at low grate speeds. Dry
refuse, on the other hand, was normally run at high grate speeds, and we found that at low
grate speeds most of the combustion was completed on the first and second grates, with
little combustion taking place on the burnout grate. However, it was still possible to vary
feed rates for dry refuse between 35 ft/hr and 55 ft/hr with only minor interdependences.
Air Flow Rates and Distributions
Because of the natural tendency toward a constant furnace air flow induced by the
natural-draft stack, the effects of varying air flows and distributions were difficult to detect.
However, various air distributions are used primarily in the underfire air system where the
air flow rate is related to burning rates. Since most of the leakage is into the overfire region,
these effects were relatively unaffected.
The underfire air can be divided into six discrete streams, since each underfire air
plenum has two sections. In practice, controlling dampers become clogged with ash reducing
the effectiveness of control. Care was taken during tests to ensure that these dampers did
operate effectively, but some reduction of control was still observed.
The quantity of air flowing under each grate was measured by means of a pitot tube.
The flow was initially calibrated by complete traverses of each duct, and periodic checks
were made to verify the continued accuracy of the sample point readings. While it was
hoped that the underfire air would be in excess of stoichiometric air, this could not be
readily achieved, particularly under the first two grates.
52
-------
The distribution of air was not as controllable as desired either. Only very small
amounts of air (5% stoichiometric) could be forced under the drying grate without causing
the draft to drop to zero (or slightly positive pressure), forcing smoke and fire out of the
furnace. The second grate underfire air generally accounted for 15 to 30% of the stoichio-
metric requirement, while the remaining 35-40% of stoichiometric air fed under the grate
came through the third or burnout grate. Because of a combination of buoyancy effects
(larger drafts at lower elevations and a lower pressure drop through the residue of the
burnout grate), it was difficult to reduce the air under that grate significantly. In addition, a
large amount of leakage was present around this end of the furnace.
The two existing sources of controlled overfire air are the roof jets, used primarily for
furnace cooling, and the wall-cooling air. During most tests the roof jets were off, and the
rate of the wall-cooling air was consistently about 4,000 cfm.
The overfire jet system described earlier could supply up to an additional 20,000 cfm.
It is interesting to note that at high jet flow rates the leakage was noticeably curtailed. The
various types of flow distributions common to the jet system are discussed in detail in the
section covering test results (see Section VII).
The remaining flow was uncontrolled resulting from the natural tendency of the
furnace to draw air from the outside. The leakage was great typically between 20,000 and
30,000 cfm, but under the right conditions this could be reduced to a very low flow rate. In
fact, very high forced air flow rates would result in almost zero leakage, but would effect a
noticeable increase in the amount of smoke emissions onto the operating floor. In addition,
the reduction in leakage usually resulted in an increase in the nose temperature of the grates
(loss of natural cooling). These factors effectively limited the amount of underfire air that
could be used.
Furnace Residence Time
Each of the three grates can be controlled independently so that the residence time for
the refuse is simply the sum of the residence times on each of the three grates. Under most
runs the residence time was approximately one hour. Certain combinations of grate speeds,
however, are not acceptable, as discussed previously, because of the tendency of the drying
grate to overheat. For most runs the three grate speeds were set at the same speed, but some
tests were made at relative residence times both slower and faster than the normal
condition.
Test Sequence
In each test the run conditions were first established and the incinerator was allowed to
reach steady-state operation before test data were taken. During the first 30 minutes to one
hour, the various parameters were monitored and final checks were made on the instrument-
ation to ensure that no major difficulties would be encountered during the run. This waiting
53
-------
period was necessary to allow the furnace to reach steady-state operation at the test
conditions, and was approximately equal to residence time of refuse in the incinerator.
Although two or three residence times might be desirable from a theoretical standpoint, the
large variation in the refuse composition over the course of a run was enough to obscure
minor deviations from the true steady state and additional waiting time was not warranted.
In the initial sets of tests used for verifying the analytical methods, data were taken over a
period of 4 to 5 hours. Because of improvements in the instrumentation and analytical
procedures, most of the data for the jet tests were taken over a period of one-half hour on
most runs. In this time two complete sweeps of the three breeching zones could be achieved,
corresponding to two sequential 15-minute average moisture determinations. The gas analy-
sis was recorded before and after the test and trends were compared to assure that the gas
averages were reasonably representative of the overall operation. The results of these tests
are presented in Section VII.
Input Data
During the course of each test the following data were recorded:
o The furnace control panel readings, which include the furnace breech
temperature (high temperature), the stack temperature, the furnace draft
pressure, the chimney draft pressure, the forced-draft fan pressure, the
under-grate pressure, and the grate speed for each of the three grates;
o The air distribution and flow rate, including velocity measurements for each
of the three underfire ducts and the overfire duct;
o The velocity of each jet, as well as the total flow to the fan and to each jet
manifold;
o The static pressure underneath each grate (during selected runs), and
o Overall furnace conditions and specific events (taken chronologically) which
might affect the test data or furnace performance, including any shutdowns,
changes in refuse feedstock or furnace operating conditions, the degree of
burnout, and the approximate condition of the refuse feed.
The emissions data and the data reduction techniques used are discussed below.
Emissions Data
Measurements of emission composition, temperature, and flow were made in the
breech and over the bed. Further measurements of static pressures were made at selected
points to allow estimation of air leakage into the incinerator. Particulate emissions were
made in the stack during the initial baseline series of tests and are reported in detail in
Section VII.
54
-------
Output flows were calculated from the breeching area velocity data collected using
S-type pitot tubes on the gas-sampling probes and the temperature measurements. Com-
position data were obtained to measure the overall combustion characteristics and the
combustible emissions. The species selected for measurements were:
CO2 - carbon dioxide H2 - hydrogen
O2 oxygen CO carbon monoxide
N2 nitrogen HC total volatile and gaseous hydrocarbons
H2 O - water
In addition, during the particulate measurement samples were collected for
Cl chloride (as hydrogen chloride),
SOX - total sulfur oxides, and
NOX total nitrogen oxides.
The collected particulate was also analyzed for trace metals.
Measurements were made at several points in the incinerator system. A cross section of
the incinerator is shown in Figure 13. The majority of the measurements were made in the
breeching at points D, E, and F. During studies of the bed-burning model, composition
measurements were obtained along the bed at points A, B, and C. The continuous-stack CO
monitor drew samples from point H. Initially velocity measurements were made only at the
points of gas composition measurement, but in the final jet-mixing phase of the program,
velocity and pressure measurements were made at points A-H as discussed later.
Initial experimental design called for operating the incinerator for an entire day
(5-6 hours) at a single condition to normalize refuse-variation effects. On that basis the
analysis system was designed on a time-sharing basis so that a single set of analytical
instruments could be used to measure the composition of three points in the breeching or
furnace. Velocity and pressure measurements were also time-shared on a single instrument,
while temperatures were all recorded simultaneously on a multipoint recorder. Problems
with obtaining accurate moisture analyses necessitated the use of an independent sampling
line piggybacked on the gas probe for H2O determinations. The general overall measurement
approach is represented schematically in Figure 14.
Analytical Procedures
The basic sampling systems are shown schematically in Figure 15. There were three
such identical systems connected to the gas analysis instruments on a time-shared basis. The
probes, cyclones, and H2 O silica traps were mounted on the staging at the breeching (or by
the furnace). A shed approximately 3x3x4 meters was constructed on the building floor
beneath the breeching, and the gas, pressure, and temperature lines were brought to the
equipment in the shed. There was approximately 10 m of sample line between the probes
and the shed.
55
-------
TEMP.
H20
TEMP.
H20
TEMP.
H20
ANALYTICAL INSTRUMENTATION GASES, PRESSURE (API
Figure 14 Overall emissions measurements approach.
STAGING
ANALYTICAL SHED
6 Ipm
BALLAST
(45min)
(LATER REMOVED)
BALLAST
(15 min)
(LATER REMOVED)
Figure 15 Sampling system.
56
-------
The original intent was to make possible the collection of the H2O for analysis by
condensation and volume measurement at the end of a run. Similarly, particulate would be
collected in the cyclone, lines, knockout flasks, and first filter. Our initial runs showed that
the moisture content was quite low, and that only about one-half of the H2O could be
collected by condensation. Consequently, a separate sample line not cooled was
strapped to the water-cooled gas sample probe and H2O determined by weight pickup on
1 cm x 10 cm silica gel traps connected to a conventional gas-sampling system.
The three probes sampled gas constantly at a high enough flow rate so that we could
always be assured of a representative incinerator sample, even though there was a relatively
large volume in the system. The combination of a lower than desired flow rate required to
prevent too high a pressure buildup on the 10-cm filter and the probe wall-effect problems
eliminated the possibility of obtaining representative particulate samples for %C analyses.
Smaller pumps constantly pulled an aliquot of the gas from each of the three filtered
combustion gas streams for the instrumental measurements. In the baseline series of tests,
gas chromatographs (GC) were used for the H2, CO2-O2-N2 analyses. One of the GC units
required 15 minutes for a full cycle; therefore the portion of the gas stream (0.1 Cpm) used
for these analyses was passed through a 4-£ container having a 45-minute time constant. In
this manner the analyzers could switch between the gas stream every 15 minutes and obtain
a sample for analysis representative of the entire time period preceding the last analysis of
that line. The more rapid CO and HC analyzers were cycled every 15 minutes, 5 minutes per
sample line.
The results from the baseline studies showed that incinerator conditions really could
not be held constant for 5 to 6 hours, and that observations would have to be made over a
shorter time (1 hour). Consequently, the ballast cans in the gas sample lines were removed
and continuous CO2 and O2 analyzers were substituted for the GC analyzer. Each sample
line was then analyzed for 5 minutes, giving a 15-minute cycle time for all three probes for
each analysis.
The gas analysis system is shown schematically in Figure 16 as used in the preferred
configuration with continuous analyzers and without the ballast cans. The much greater
time variation in composition than expected places a clear preference on continuous and
rapid analysis, and we strongly recommend this approach over the slower batch-type
analyses in all future studies.
Basically the instruments are simply switched between the three probe gas lines by
means of the solenoids SI-S3 and S'l-S'3 and cycled between them every 15 minutes with 5
minutes on each probe line. The two-way solenoids allow continuous flow through the lines,
so that a representative sample of the incinerator at that moment is always present.
The calibration gases were manually switched into the instruments for zero and span
checks at the beginning and end of each period of data collection.
57
-------
PROBE LINES
oo
CALIBRATION
GASES
HiCO2,O2
L0C02,02
ZERO AIR
PROPANE
1
>v
> 1
v
S1
\ v
1
V
j
a
) S2
^ V
) (
S'1
k /
S (
S'2
X.-..I
ANALYTICAL PUMPS
SOLENOID
TIMING
CIRCUIT
^TS
S'3 ^.1 X
Figure 16 Gas analysis system
-------
The starch CO/CO2 monitor ran continuously during periods of test activity monitor-
ing the concentration at point H in Figure 13.
The pressure measurements were made with the S-type pitots on the gas-sampling
probes and various other static pressure taps in the incinerator system, as indicated in
Figure 13. The pressure-measuring lines were connected through a separate timer-controlled,
solenoid switching manifold system, as shown schematically in Figure 17. A set of 10
measurements was made and recorded automatically every 10 minutes. The sequence of
measurements is given in Table 3. The recorded signal from the capacitance manometer was
calibrated for ± 1 cm H2O with an inclined manometer.
Table 3. VELOCITY AND STATIC PRESSURE MEASUREMENT LOCATIONS
Sequence Function Probe
Probe Location'
1
2
3
4
5
6
7
8
9
10
Manometer Zero Check
Velocity S-Pitot
Static Pressure* Static Tap
Velocity S-Pitot
Velocity S-Pitot
Static Pressure* Static Tap
Static Pressure* Static Tap
Static Pressure* Static Tap
Pressure Drop across Scrubber Static Taps
Furnace Room Outside Building Static Taps
Pressure Comparison
D
D
E
F
F
B or A**
C
GandH
Furnace Room
and Outside Building
'Reference: furnace room pressure
**Location changed from B to A for measurements on and after April 17, 1973.
t See Figure 13.
ANALYTICAL EQUIPMENT
Probe
Considerable initial evaluation and trade-off consideration went into the final design of
the gas-sampling probe. The first experiments showed that an uncooled 1.2-cm SS tube
could not physically withstand the 900-1000°C breeching temperatures and, more impor-
tantly, that further reaction of the gases occurred in the hot sampling tube. When evaluated
59
-------
TO PITOTS AND STATIC PRESSURE TAPE
o\
o
TO SOLENOID VALVES
FUNCTION
INDICATOR
LIGHT
PANEL
TIMER
(CAM TYPE)
4 cj> 4
4
SOLENOID
VALVES
PURGE LINES
_Q (TO COMPRESSED AIR TANK)
Q Q MANUALSHUTOFFVALVES
ZEROING SOLENOID VALVE
CAPACITANCE
MANOMETER
STRIP
CHART
RECORDER
Figure 17 Static pressure and velocity (AP) system
-------
materials which might physically withstand the temperatures either had an excessively high
cost or a high probability of breakage. The EPA loaned us a water-cooled probe* for design
evaluation, and we found that, using this probe, we could quench the combustion gas
immediately after it entered the probe, and that relatively inexpensive materials could
survive the incinerator environment without too much damage.
Three probes were designed and constructed as shown schematically in Figure 18
(Section AA). This type probe essentially consists of a pair of concentric tubes with an inner
tube for inlet cooling water. An additional internal tube was included for transmission of
the thermocouple leads and fastening the thermocouple. A glass liner was used to decrease
catalyzed wall reactions of the combustion gases. The first probes had a 90-deg sample inlet
so that samples, including particulate, could be obtained parallel to the flow of the gas
stream. Two problems were encountered with this design. The welds in the region of the
90-deg angle could not withstand prolonged 1000°C temperatures. In fact two probes failed.
Furthermore, we found that the impacting fly ash simply caked over the gas inlet and
caused sampling problems. Consequently, the ends of the probe were cut off at Section AA
and reconstructed as shown in Section BB. Sampling perpendicular to the gas stream
eliminated the fly ash caking problem and none of the modified probes failed.
The S-type pitot tubes clamped to the probe worked well when clean, but, like the gas
inlet, the upstream side of the pitot constantly became caked over and fused. Reliable
pressure (AP) measurements could only be obtained with frequent (every hour) removal of
the probe and cleaning of the pitots.
It was very important to maintain a high flow of cooling water through the probes,
since even a very brief overheating caused probe sagging, linear breakage, and a run abort. In
a project such as this, even the simplest things become problems. For example, we found
that high-quality garden hose used to supply the probes could not be trusted and had to be
carefully pressure-checked for burst strength.
Temperature measurements were made with an Omega Pt/Pt 10% Rh thermocouple
housed in a ceramic casing. The thermocouple was held on the end of the probe at
position 1 in Figure 18 as shown in Figure 19. The details of the radiation shield used to
protect the thermocouple from the incinerator radiation and assure accurate gas tempera-
tures are also shown.
Some measurements of "temperature only" were made over the burning refuse bed,
while gas compositions were determined in the breeching. For these measurements, three
simpler water-cooled probes were made similar to the gas probes, but they contained only
the line for the thermocouple and the radiation shield.
"Courtesy of Mr. James Dorsey, Control Systems Laboratory.
61
-------
ON
to
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
.LI
co co
m m
O O
m
THERMOCOUPLE WELL - 3/8-INCH BWG20, 316 S.S.
COOLING WATER - 1/4-INCH TYPE L, COPPER TUBING
SAMPLE OUTLET - PYREX GLASS LINER, 5/8-INCH O.D.
16 M.M., WITH 28/12 BALL JOINT
SAMPLE TUBE RECEIVER - 1/2-INCH SCHEDULE 10, 316 ELC-S.S.
SAMPLE INLET - 3/8-INCH SCHEDULE 80, 316 ELC-S.S.
S-TYPE PITOT TUBE - 3/8-INCH BWG 20, 316 S.S.
COOLING WATER INLET - 1/4-INCH SCHEDULE 40, 316 ELC-S.S.
COOLING WATER OUTLET - 1/4-INCH SCHEDULE 40, 316 ELC-S.S.
PROBE JACKET END PLATE - 316 ELC-S.S., 11 GAGE 2B FINISH
PROBE JACKET - 1 1/4-INCH SCHEDULE 10, 316 ELC-S.S.
MOISTURE SAMPLING TUBE- 1/4-INCH BWG20, 310S.S.
WORM-DRIVEN HOSE CLAMPS
co r= H oj
E! r- o m
Z S 2 ""
£°>53
10 I > CO (/)
l>5°o
I O O 5 o
sgs§3
lll|i
3 o H 6 3
I CO Z O
sl°;
r= I o 3J
H > oo >
= H m Z
S co <. O
5>ls
2l3l
m .'
m
Figure 18 Water-cooled sampling probe.
-------
21/20.D.x0.049,304S.S,
1 1/2 O.D.x 0.049, 304 S.S,
_0.035 D.
304 S.S. WIRE
3/8 O.D. x 0.035 WALL, 304 S.S.
FULL WELD
316 S.S.
.SET SCREW -2 PLACES, 90° APART
3/4" D
. END OF COOLED PROBE
1/2 O.D.x 0.035, 304 S.S.
MODIFICATIONS
2 1/2 O.D. WALL INCREASED TO 0.065
1 1/2 O.D. WALL INCREASED TO 0.065
3/8 O.D. TUBE REPLACED WITH 1/4 INCH SCH 40, 316 S.S.
1/2 O.D. TUBE OMITTED
0.035 INCH WIRE REPLACED WITH 0.093 316 ELC S.S. ROD
WEDGE WAS ADDED FOR STRUCTURAL SUPPORT, 2 x 1 90° TRIANGLE
MATERIAL-316 ELC S.S., 11 GAGE
Figure 19 Thermocouple radiation shield
63
-------
Instrumentation
The following instrumentation was used for monitoring the concentration of the
species indicated:
H2 Carle Model 8004 gas chromatograph with automatic timer and sample
injection.
CO2,O2 - Carle Model 8004 gas chromatograph with automatic timer, sample, and
N2, CO column switch.
CO2 - Beckman Model 31 SB NDIR.
- MSA Model 300 LIRA NDIR.
CO - MSA Model 300 LIRA NDIR (two).
HC - Beckman Model 108A FID.
O2 - Beckman Model 742 polarograph.
Pressure MKS Baratron capacitance gauge.
Temperature L and N multipoint recorder.
The main gas-probe sampling pumps were made by Cast (Models 0322), and the
analytical measurement pumps were made by Metal Bellows Corp. (MB-21). Rockwell dry
gas meters were also used, and RAC inline aluminum filters were used for both locations. All
fittings were Swagelok, and Nupro needle and toggle valves were used. All components of
the gas analysis system were brass and copper.
Photographs of Equipment
Figures 20 through 23 are photographs of the various pieces of equipment in the
analytical shed. Figure 20 is the three main gas sampling systems as represented earlier in
Figure 15. Figure 21 shows the multipoint temperature recorder and the original three-pitot
pressure-measurement system. The pressure system was later expanded to the 10-point
measurement manifold. Figure 22 is a view including the NDIR CO and FID HC analyzers
and several of the recorders. Figure 23 shows the two original gas chromatograph analyzers
(one was later replaced with continuous CO2 and O2 analyzers) and the CO NDIR
analyzers.
64
-------
Figure 20 Gas sampling lines, pumps and meters.
-------
Figure 21 Recording temperature and pressure systems.
66
-------
Figure 22 NDIR and FID analyzers and recorders.
67
-------
Figure 23 Gas chromatograph and NDIR analyzers.
68
-------
DATA REDUCTION
The data reduction was primarily a three-step operation. The form of the data was first
changed from percent displacement on a strip chart recorder to a tabular listing of data in
appropriate units. In the second step, the data for a particular run were averaged so that each
run could be characterized by a simple set of output parameters. Finally, the validity of the
output parameters was checked by mass and energy balances.
Data Recording and Averaging
The first and second steps were purely mechanical manipulations of the recorded data.
In reading data points from a continuous strip chart recorder, five-minute averages were
taken at five-minute intervals. This particular method was chosen because of the five-minute
sample time for each of the breech zones. The raw averaged data were then applied to the
appropriate calibration curve to establish the data in desired units of temperature, concen-
tration, or velocity. The moisture content data were recorded in 15-minute intervals as
15-minute averages; this timing corresponded to one complete cycle of the breech zone
sampling system. To calculate the set of output parameters for each run, the tabulated data
were time-averaged for the entire run. The output parameters are listed below:
Temperature
Breech zones 1, 2, and 3 and the average temperature over the entire breech
Flow rate
Velocity in breech zones 1,2, and 3
Total molar flow rate
Composition of exhaust gas
Breech zones 1, 2, and 3 - CO2, CO, CH4, H2O, and 02, and the average
molecular weight
For total breech - CO, CH4, O2, N2, CO2, and H2 O
Heat output of system
These output parameters were used for the remainder of the analysis.
Data Checks Mass and Energy Balances
The difficulties encountered in obtaining valid data from an operating municipal
incinerator are well known. The conditions under which sampling must take place are hot
(1200 to 2000°F), dirty (molten fly ash, heavy organics, high moisture content), and not
chemically constant from second to second or, on the average, from hour to hour. These
difficulties do not preclude the possibility of obtaining valid data, but they do place
increased significance on material and energy balances as a means of verifying the internal
69
-------
consistency of the data. Although we accept that mass and energy balances are not the
necessary and sufficient criteria for proving the validity of data, reasonably good balances
do tend to increase the credibility of the data and, hence, our confidence in it.
As a rule, then, whenever balances can be closed, the data are considered valid. For
experiments in which the balances did not close, the data were rejected. In the latter cases,
we attempted to explain the most probable source of error or inconsistency.
The Mass Balances
The overall mass balance over the incinerator was impossible to obtain because a large
fraction of the air entering the furnace was not metered (leakage). The amount of leakage
was determined by the difference between the exiting gases and the forced air and
combustion-generated gases.
Carbon The only important source of carbon in the system is the refuse feed. It is
assumed that the average refuse at Newton is 50% combustible, and that these combustibles
have an empirical formula of CHj 585O0 675 based upon previous data for Newton [8].
Carbon leaves the system as CO2, CO, various hydrocarbons, soot, and unburned material
which falls out of the incinerator as ash. For a typical run the distribution of carbon among
these various forms is shown in Table 4.
Table 4. CARBON DISTRIBUTION PROFILE
Component
Pounds
(mols C/hr)
Percent
Carbon In
Refuse
CO2 in air
415.71
413.35
2.36
100
99.43
0.57
Carbon Out
C02
CO
Hydrocarbons
Soot
Ash
415.71
410.31
1.49
0.56
0.21*
4.14*
100
98.46
0.36
0.13
0.05
1.00
'Arthur D. Little, Inc., estimates.
70
-------
Because carbon dioxide represents such a large percentage of the total "carbon out," in
the remainder of the analysis we have ignored the contributions from CO, hydrocarbons,
soot, and the ash. By the same reasoning only the "carbon in" as refuse will be utilized in
further carbon balances.
For these specific data, the combustion of refuse is represented by the following formula:
CH1.585 °o.625 0.655 H2O+ 1.084 02 +4.078 N2>CO^ + 1.448 H2O+ 4.078 N2
(44)
This relationship suggests that 1 pound mol of CO2 should be produced for every 23.585
pounds of combustible (or 47.17 pounds of refuse). Because of the uncertainty and
variations in the bulk density of refuse, as well as the voids which are present in the feed
hopper, it is difficult to calculate the exact feed rate. However, for the purposes of the
material balances, feed rates were calculated assuming a density of 13 lb/ft3 and that there
were no voids in the feed hopper. It can be seen from Table 4 that the carbon in and the
carbon out are in reasonable agreement. From the information in this table one can
calculate that the average amount of combustible fed into the furnace was 51.5% of the
refuse (p = 13 lb/ft3). For the purposes of the current work, this agreement is adequate to
verify that the techniques and apparatus are appropriate.
Hydrogen Hydrogen enters the system as water vapor in the air, as moisture in the
refuse, as bound hydrogen in the refuse, and as water evaporated from the ash quench
system. The hydrogen leaves the system as water vapor and hydrocarbons in the exhaust
gases and as moisture and bound hydrogen in the unburned combustibles. A hydrogen
balance for a typical run is shown in Table 5. As in the case of the carbon balance, the small
contributors have been ignored; that is, hydrocarbons and hydrogen in ash either bound or
as moisture and the water vapor in the air from the ash quench system. The "hydrogen out"
is all measured as water vapor. The "hydrogen in" is calculated from the "carbon dioxide
out" (which has previously been related to amount of refuse fed). The empirical relationship
states that the mols of water released is equal to 1.448 times the mols of CO2 released. As
can be seen, the water balance (which implies the hydrogen balance) is good enough to
justify the use of the data. In the current study we found the average moisture content of
refuse to be 24.6%, as compared to the assumed value of 25%.
71
-------
Table 5. HYDROGEN BALANCE FOR TYPICAL RUN
Pound
Component (mols H/hr) Percent
Hydrogen In 1373.24 100
HasH2OinAir 122.90 8.94
H as H20 in Refuse 541.58 39.45
H (bound) in Refuse 655.16 47.71
H as H20 from Ash Quench* 53.60 3.9
Hydrogen Out 1373.24 100
H20 in Gases 1359.13 98.97
H as Hydrocarbons 2.13 0.16
H20 in Ash 5.42 0.39
H (bound) in Ash 6.56 0.48
'Arthur D. Little, Inc., estimate.
Oxygen The oxygen in the system comes from the atmospheric oxygen and water
vapor, the moisture in the refuse, the water evaporated from the ash quench, and the bound
oxygen in the refuse combustibles. The oxygen leaves the incinerator as atmospheric
oxygen, carbon dioxide, carbon monoxide, and water vapor. A typical run has an oxygen
balance similar to that presented in Table 6.
Table 6. OXYGEN BALANCE IN TYPICAL RUN
Pound
Component (molsO/hr) Percent
Oxygen In 3673.65 100
O as O2 in Air 3056.27 83.20
OasH2OinAir 61.45 1.67
0 as H20 in Refuse 270.79 7.37
0 as H2O from Ash Quench* 26.80 0.73
0 as Bound Oxygen in Refuse 258.34 7.03
Oxygen Out 3673.65 100
O as O2 in Gases 2166.68 58.98
OasH2Oin Gases 679.56 18.50
0 as C02 in Gases 820.62 22.34
0 as CO in Gases 1.49 0.04
OasH2OinAsh* 2.71 0.07
O as Bound Oxygen in Ash* 2.59 0.07
'Arthur D. Little, Inc., estimate.
72
-------
The small contributors - CO, oxygen in ash as moisture and bound water, and water from
ash quench are not considered when solving the oxygen material balance.
A summary of the average oxygen balances is given in Table 6. As can be seen, these
balances are adequate for the current work.
The Energy Balance
The energy balance over the incinerator has been used in the same way as the mass
balances to verify the experimental techniques and the analytical apparatus. Essentially,
there are two ways of conducting the energy balance. The first is to assume a constant
heating value for refuse and check the heat actually liberated with the theoretical heat
available. The other is to compare the heat liberated per unit of combustion product
formed. The latter is the chosen course because it does not presume knowledge of the heat
content of the raw refuse feed or the feed rate.
It is known that about 5650 Btu's are released per pound mol of CO2 liberated in the
combustion of many organic materials. Therefore, one could conduct the heat balance by
knowing the mass flow, composition, temperature, and heat capacity of the gas stream. On the
other hand, the same heat balance could be accomplished knowing only the fraction of the
flue gas which is CO2 and the temperature. If we assume that the heat capacity of the flue
gas stream does not change with composition, then the temperature of the flue gases will be
a linear function of the fraction of CO2 (Appendix D, Eq.D-31). Furthermore, the theoretical
line passes through the adiabatic flame temperature at the stoichiometric CO2 concentration
and through ambient temperature at a CO2 concentration equal to zero. Figure 24 shows
the data on such a plot of temperature vs CO2 concentration. Because the system is not
adiabatic, it is reasonable to assume that the dotted curve represents the kind of deviation
from the theoretical that we might expect, that is, the largest deviation at higher tempera-
tures and no deviation at ambient. Rotation of the line about the average temperature
would also be expected as a result of radiation. From the slope of a line through the plotted
points, which is about .004% C02/°F (or 260°F/%CO2), we have calculated that the heating
value of the refuse is about 4650 Btu/lb based upon Eq. (D-31) of Appendix D.
73
-------
1 -
750
1000
1250
TEMP.,°F
1500
O ZONE 1
A ZONE 2
D ZONE 3
1750
2000
Figure 24 CO2 vs. temperature.
74
-------
SECTION VII
TEST RESULTS
The strategy we espoused for controlling combustible emissions in Section V was
predicated on two separate mixing concepts:
Macroscale mixing to achieve uniformity throughout the furnace; and
Microscale mixing to promote burnout of combustibles at the molecular
level.
The equations that we presented as the basis for each of the two mixing models are
verified in this section. The data were taken at an existing municipal incinerator which was
retrofitted with overfire air jets for this program.
The test equipment and procedures were discussed in Section VI. In this Section, we
will describe the specific tests that we carried out and discuss the results of those tests. A
compilation of the test data is given in Appendix A.
MACROSCALE MIXING MODEL
Overfire jets can be used to achieve uniform mixing throughout the furnace. The
macroscale model proposes a design basis for overfire air jets based upon:
Ivanov's equation for jet penetration, and
Use of Bernoulli's equation to characterize the flow patterns within the
furnace.
The method for calculating the penetration of an overtired jet into the furnace was
suggested by Ivanov and has been reviewed in Section V. The design equation (Eq. (3)) is
repeated below:
i
d:Uj
L-L (3)
The equation is dependent upon the jet diameter, its velocity, and the crossflow
velocity of gas in the furnace. In a series of jet tests the penetration distance was estimated
visually and was compared with the calculated value for penetration. Figure 25 shows a plot
of penetration vs jet velocity for 5-inch diameter jets firing into a crossflow velocity field of
15 ft/sec. The crossflow velocity was calculated using the fluid-flow equations discussed
previously. Clearly, the Ivanov relation is adequate for estimating jet penetration, and the
deviation from the theoretical relationship is well within experimental error.
75
-------
o
{= t; 4
tJ"
dJ
= 1.6d,u,
= 5"
= 15fps
p\
= 1-8
10
20
30
40
50
60
70
80
90
100
AVERAGE JET VELOCITY, Uj.ft/sec
Figure 25 Penetration vs. jet velocity.
FLU ID-FLOW PATTERNS
The fluid-flow model developed in the mixing study is based on the premise that
Bernoulli's equation can be applied along the streamline of flow in an incinerator so that
both the residence time and the crossflow velocity at the level of the jets can be calculated.
During the tests, velocity and temperature were measured in the breech of the
incinerator. The profiles clearly showed the stability of the hot zone at the top of the
breech, but the data did not show large difference between the "hot" and "cold" zone
velocities. The equations derived for a sealed furnace [Eq. (8)] could not correlate the data,
as witnessed by the random appearance of a plot of u2 vs 2g(z-z0) [(TH/TC) -1 ]. On the
other hand, Eq. (9) derived for unsealed furnaces could correlate the data reasonably
well. A plot of u2 vs [(T/T0)-1] where T0 is the ambient temperature and Tis the
76
-------
temperature of either the hot zone or cold zone is shown in Figure 26. The plot should be
linear if Eq. (9) repeated here is applicable to incinerators:
Hot Zone , v Cold Zone
With the exception of a few spurious points, the agreement with the theoretical
relationship is excellent. The solid line in the figure represents the best fit through the data
points corresponding to a furnace height of 12 feet and an initial velocity of 18.4 ft/sec. The
dashed line represents the curve passing through the origin, so that the furnace height is 13.7
feet, while the initial velocity is zero. Because of the gross nature of the flow patterns within
the furnace, the analytical distinction between the height of 12 feet and 13.7 feet is not
possible. Either curve could be justified as confirmation of the Bernoulli approach; neither is
so well defined that it may be used to predict initial velocities.
Two conclusions can be drawn from this analysis:
1) Bernoulli's equation can, indeed, be used for design purposes in determining
exit velocities and approximate furnace residence times; and
2) In a furnace which has considerable air leakage, both the hot and cold zones
will accelerate according to the Bernoulli relation. The driving force for the
acceleration is the buoyancy of gases within the furnace as compared to
outside the furnace.
Verification of Eq. (10) of Section V, which describes the difference in velocity of the
hot zone vs the cold zone, was not possible. The accuracy of the data does not warrant such
fine tuning of the fluid-flow model.
MICROSCALE MIXING MODEL
The theories presented to characterize the microscale mixing introduced the use of
Hottel's unmixedness factor and Corrsin's equation as fundamental relationships charac-
terizing the three T's on a microscopic scale. Hottel's and Corrsin's concepts have been
verified in controlled situations, but because of the large variations in incineration data it is
unrealistic to hope for conclusive evidence that will provide exact quantitative relations. The
data are sufficiently accurate, however, to demonstrate the consistency of the data with the
two concepts.
The Unmixedness Factor
In Section V, the relationship between the unmixedness factor suggested by Hottel (4)
and the ratio $ of available oxygen to oxygen required to complete combustion was
77
-------
oo
3.0
where
T.=70 F, us 18.4 ft /sec
a LI
z-z0 =12ft.
Tas70°F, u0 =
-z)^ 13.7ft.
1.0
1500
1750 2000 2250
u,2 [ft2/sec2]
Figure 26 Confirmation of Bernoulli's equation.
2500
2750
3000
-------
presented. The relation is derived in Appendix F. The procedure for calculating the value of
the unmixedness factor is outlined in Appendix B. Supporting data required to characterize
the refuse bed offgas is presented in Appendix E.
The unmixedness factor can be calculated for each gas sample that is taken. Its value is
dependent only upon the gas analyses and the empirically determined value for CO/CO2. The
experimental value for yV^/Ay is determined from Figure 3, based on the value of $
[(available oxygen)/(oxygen required to complete_combustion)], determined from the gas
analysis. The value of the unmixedness factor, y c'2, is calculated from the following equa-
tion which is based upon Eq. (B-12):
77 \ / o/*r*r\ \
(45)
In Appendix B, the quantity ys is shown to be equal to 0.78 when CO/CO2 is equal to
0.5. The composition of the combustibles and combustion products (without entrained air)
is also given.
In Table 7, the experimental value for y c'2 is given for each breech zone and each of
the 28 jet tests.
To compare the experimental value for unmixedness factor with the theoretical one,
the following data are required:
original CO/CO2 off the bed, and
relationship between temperature and CO2 in the breech.
In these calculations, a value of 0.5 for CO/CO2 has been assumed.
The plot of temperature vs %CO2 |w was given previously in Figure 24.
The data can be represented by the linear relationship:
%CO I
2 w =(Temp-500°F)/1500°F (46)
10%
The calculation procedure is as follows:
1. For a given temperature, calculate %CO2 |w from the equation given above;
2.
^
3. For a given value of yc'2 , calculate lye'2 /^yy ;
79
-------
Table 7. SUMMARY OF UNMIXEDNESS FACTORS
Zone I
Zonal!
Zone III
Run#
22-1
222
22-3
23-1
23-2
23-3
23-4
24-1
24-2
24-3
24-4
OO 25-1
° 25-2
25-3
25-4
31-1
31-2
31-3
32-1
32-2
32-3
33-1
33-2
33-3
34-1
34-2
34-3
34-4
Emission
Factor
1
1
1
127
87
330
252
144
60
55
409
652
976
106
91
157
46
26
243
207
235
275
86
505
272
67
61
54
Temp.
1721
1508
1700
1715
1437
1717
1421
1647
1694
1551
1730
1900
1796
1400
1523
1750
1507
1525
1825
1683
1698
1777
1666
1753
1793
1554
1497
1597
V^L
<0.12
<0.14
<0.12
0.145
0.165
0.160
0.190
0.150
0.135
0.150
0.165
0.155
0.180
0.170
0.160
0.145
0.170
0.145
0.140
0.155
0.155
0.150
0.140
0.150
0.150
0.150
0.155
0.140
V^w
.062
.076
.065
0.116
0.157
0.149
0.200
0.174
0.115
0.126
0.132
0.099
0.193
0.147
0.135
0.093
0.135
0.125
0.139
0.130
0.161
0.113
0.130
0.159
0.138
0.129
0.138
0.118
^
_
_
-
0.029
0.008
0.011
- 0.010
- 0.024
0.020
0.024
0.033
0.056
- 0.013
0.023
0.025
0.052
0.035
0.020
0.001
0.025
- 0.006
0.037
0.010
- 0.009
0.012
0.021
0.017
0.022
Emission
Factor
1015
583
244
677
909
3705
2707
535
18
24
24
499
462
568
1387
267
442
572
49
182
418
550
416
545
320
502
596
412
Temp.
1446
1387
1458
1446
1125
1350
1158
1385
1470
1341
1426
1696
1587
1246
1131
1543
1261
1308
1707
1667
1609
1545
1485
1585
1478
1417
1367
1481
V^
0.222
0.210
0.180
0.210
0.250
0.290
0.290
0.210
0.150
0.160
0.150
0.170
0.185
0.225
0.265
0.175
0.215
0.220
0.130
0.155
0.180
0.190
0.190
0.190
0.185
0.205
0.215
0.190
^u-
0.237
0.223
0.172
0.223
0.244
0.312
0.287
0.210
0.149
0.159
0.160
0.167
0.191
0.219
0.265
6.173
0.221
0.214
0.139
0.157
0.198
0.202
0.193
0.195
0.185
0.214
0.230
0.196
&
- 0.015
- 0.013
0.008
- 0.013
0.006
- 0.022
0.003
0
0.001
0.001
- 0.010
0.003
- 0.006
0.006
0
0.002
- 0.006
0.006
- 0.009
- 0.002
- 0.018
- 0.012
- 0.003
- 0.005
0
- 0.009
- 0.015
- 0.006
Emission
Factor
1314
274
931
460
979
12.162
7953
1102
1633
2461
2508
823
1242
2060
5141
823
193
142
1403
84
1042
2449
2012
3340
2484
2594
2862
3689
Temp.
1029
1004
1100
1046
737
871
856
892
1152
949
1071
1283
1225
875
742
1193
836
1025
1381
1408
1344
1251
1240
1306
1150
1106
1023
1115
1C '«xpt.
0.270
0.225
0.250
0.235
0.280
70.40
0.370
0.270
0.270
0.300
0.295
0.230
0.250
0.295
0.350
0.240
0.230
0.210
0.240
0.170
0.230
0.280
0.270
0.290
0.290
0.295
0.305
0.320
v^
- 0.280
0.240
0.278
0.239
0.241
0.350
0.338
0.262
0.296
0.306
0.308
0.304
0.270
0.297
0.325
0.265
0.241
0.231
0.290
0.211
0.272
0.304
0.293
0.311
0322
0.312
0.315
0.325
^
- 0.010
- 0.015
- 0.028
- 0.004
0.039
0.050
0.032
0.008
- 0.026
- 0.006
- 0.013
- 0.074
- 0.020
- 0.002
0.025
- 0.025
- 0.01 1
- 0.021
-0.050
- 0.041
- 0.042
- 0.024
- 0.023
- 0.021
- 0.032
- 0.017
- 0.010
- 0.005
-------
4. Determine $ from Figure 3 ;
5. Calculate %O2 from the material balance equation (Eq. D-26 of Appen-
dix D):
%o2 t %co2i _
21% 15.1%
6. Calculate O2 required, according to the relation:
7. Calculate the emission factor according to the equation:
CO(PPm)|12% = (02req)
Note that for refuse:
%co2lw =%co2id (47)
15.1% 19.5%
Using the above steps, we calculated the data shown in Figures 9, 10, and 11. The
procedure shows that emission factors can be calculated if the temperature and
unmixedness factor are specified and if the original composition of the offgas is
known.
The theoretical value for y c'2 has been determined from Figure 8 for each of the
jet tests, and these values are tabulated in Table 7.
A careful examination of the data reveals that in all cases, even in the quench
zone, high emissions are reflected in high unmixedness factors. The agreement between
experimental and theoretical values overall is 0.003 or within 1% of the expected
value. The agreement is best in Zone II where the data from Figure 24 are best fitted
by Eq. (46).
The parameter (CO/CO2)0 has no effect on the relative values of theoretical and
experimental unmixedness factors, since the ratio was used to calculate both. However.
a change in the value of (CO/CO2) 0 results in a change in the absolute value of^/c^
Hence, to attempt to place any weight on the actual numerical values in Table 7 would
be stretching the validity of the data. With respect to the validity of the concept, the
correlation of the data is remarkably well fitted to the unmixedness factor concept.
Without this characterization factor, the more sophisticated advances in the theoretical
understanding of incinerator mixing do not readily follow-
81
-------
Corrsin's Equation
The verification of Corrsin's Equation could be based upon the defining equation,
(Eq. (21)):
(21)
or alternatively upon the comparative equation (Eq. (40)):
y
1 +
Ayb\%C02|s
d-x)
(40)
In Appendix E, tests are described wherein offgas compositions were measured directly
over the refuse bed. From these data the initial value of the unmixedness factor ^/ c7* was
calculated. The calculated values for the two complete bed tests are shown in Table 8. The
scatter in the data was so great that no meaningful conclusions were possible.
Table 8. SUMMARY OF TEST RESULTS OVER
Sample Sort
Bed Test 33
C02
02
CO
H2
HC
H20(>)
H20/C02
CO/C02
Kwg
Temperature
Bed Test 2
C02
02
CO
H2
HC
H20
H20/C02
C0/C02
^wg
Temperature
5.93
13.4
3.2
1.72
1.16
11.51
1.94
0.54
3.61
1835
13.3
6.0
3.35
2.7
1.94
12.9
0.97
0.25
1.20
1787
1.31
18.9
0.26
0.21
0.01
2.15
1.64
0.20
2.03
1905
6.4
14.2
1.00
0.77
0.34
19.4
3.03
0.16
3.94
1758
5.8
14.6
0.13
0.09
0.03
8.84
1.52
0.02
2.2
1928
12.0
8.6
0.26
0.19
0.39
26.4
1.70
0.02
2.33
1811
Average
4.35
15.63
1.20
0.67
0.40
7.50
1.72
0.28
3.1
1889
10.6
9.6
1.54
1.22
0.89
17.57
1.66
0.15
2.09
1785
1. Complete sets of data are limited due to experimental difficulties directly over the bed.
2. Calculated from material balance.
82
-------
In pursuing the alternative approach to verification of the Corrsin relation, note that
neither the parameter x nor the parameter (j3t)b are well known or clearly defined. The
value for x could be defined by Eq. (28) in Section V, based upon the ratio %CO2 |j/%C02 Ib
One might also define x based on temperature by substituting Eq. (46) into Eq. (28) of
Section V to yield:
(T: - 500)
-- (48>
Eq. (28)
0.284
0.260
0.163
0.139
Eq. (47)
0.194
0.180
0.087
0.019
Eq. (30)
0.290
0.294
0.277
0.256
Finally, one might also define x based upon flow rates (Eq. (30)). For example, the
different calculated values for x in each of the test conditions of Test 31 are shown in
Table 9 to be in relatively good agreement. The values for Test 33 are grossly apart.
Table 9. COMPARISON OF ESTIMATES FOR (1-x)
_ Basis for (1-x)
Test No.
31-2
31-3
33-1
33-3
The reasons for the discrepancies in the above values for x are:
Errors in analytical data,
Changes in total flow rates into the furnace, and
Changes in composition or heating value of refuse.
To determine the effect that different values for x might have on the resulting
calculations, consider the calculations summarized in Table 10. In these calculations,
based upon Tests 31-1 (base condition) and 31-2 (jet condition), x is chosen arbitrarily
and Ay is defined by the equation:
%C02|b
(49)
Note that as the value for (1-x) decreases, the value for the experimentally determined
unmixedness factor \/c^ also decreases (due to the change in the calculated value for
Ay, according to Eq. (49) above), and the predicted value for\/crT increases, primarily
due to the decrease in the value of (|3t)j. The result is the rapid convergence of the
experimental and predicted values for unmixedness factor to the extent that agreement
with the Corrsin relation can be shown.
In the first set of calculations (3t was assumed to be equal to 0.25. On the other
hand, the value of ()3t) is not clearly defined, because it depends on the arbitrary
83
-------
Table 10. CALCULATION RESULTS USING DIFFERENT VALUES FOR x
oo
Test
31-1
31-2
31-1
31-2
31-1
31-2
(1-x)
0
0.284
0.150
0.100
0
0.284
0.150
0.100
0
0.284
0.150
0.100
*
445
3339
3339
3339
445
3339
3339
3339
445
3339
3339
3339
From Data
f(*>
0.455
0.370
0.370
0.370
0.455
0.370
0.370
0.370
0.455
0.370
0.370
0.370
(Appendix
Ay(x)
0.207
0.371
0.294
0.265
0.207
0.371
0.294
0.265
0.207
0.361
0.294
0.265
B)
^
0.094
0.137
0.109
0.098
0.094
0.137
0.109
0.098
0.094
0.137
0.109
0.098
From Theory Application of Eq. (41 )
*0
86
203
138
119
150
354
241.
208
311
734
500
431
«*o»
0.584
0.500
0.535
0.550
0.528
0.465
0.490
0.500
0.477
0.425
0.445
0.455
V^0 1
0.121
0.186
0.157
0.146
0.109
0.173
0.144
0.133
0.099
0.154
0.131
0.121
+ ys (%C02/b) ,,._
Ay %co2/s
1.0
1.791
1.418
1.278
1.0
1.791
1.418
1.278
1.0
1.791
1.418
1.278
:) f(*j0)/f(4>b0>
1.0
0.856
0.916
0.942
1.0
0.881
0.928
0.947
1.0
0.891
0.933
0.954
*
0.250
0.745
0.420
0.321
0.150
0.447
0.252
0.193
0.050
0.149
0.084
0.064
N/c"
0.094
0.088
0.103
0.105
0.094
0.110
0.112
0.109
0.094
0.136
0.120
0.113
A%
-35.8
-5.5
+7.6
-19.7
+2.8
+11.2
-0.7
+10.1
+15.3
-------
assumption of "characteristics," lengths, and also upon assumed relationships between
residence time, furnace volume, and flow rates (Eq. (26)). In the subsequent sets of
calculations in Table 11, the effect of different values of£pt)b is demonstrated. As ()3t)
decreases, the experimentally determined value for yc7 remains constant, but the
predicted value, based upon Eq. (40), increases. At a value for (j3t)b of 0.05, the
experimental and predicted values for unmixedness factor agree (based upon x
calculated from carbon dioxide ratios).
The other variables in Eq. (40), such as the assumed values for CO/CO2 and
(Af/Aj), are much less important and have not been considered in detail.
Table 1 1 shows the relative agreement between yc determined from the stack
gas analysis (based upon Ay calculated from Eq. (49)) and the predicted value for
ye'2 , based upon Eq. (40). Note that for a given value of x, there is a value for (|3t)b
which will produce close agreement between the twoyc77 values. Likewise, for a given
value for ((3t)b , there exists a value for x which will also produce a close agreement
between the experimental and theoretically predicted values for yc72". In Figure 27, the
value of (1-x) is shown as a function of the value of (0t)b required to reconcile the test
data for Tests 31-1 and 31-2 with the values calculated from Eq. (40). The range of
mutually consistent values for (1-x) and (j3t)b falls within the accuracy of the experi-
mental data. Because of the uncertainty associated with both x and j3t, one cannot
conclude that the validity of the Corrsin theory is "proved," but only that the data are
consistent with the theory to the extent that the results of the tests could be
reasonably explained. Similar analyses could be made on each of the other series of
tests with some notable exceptions. In Test 32 the addition of jet air caused an increase
in CO2 in Zone 1 . A possible explanation for this is that the jet air was introduced at a
much higher static pressure than was the air which leaks into the furnace. The
introduction of jet air could change the static pressure pattern within the furnace to
the extent that one mole of jet air displaces more than one mole of leakage air. The
result was a decrease in the total amount of air entering the furnace and a corre-
sponding increase in the CO2 content. In this particular case then, the emissions are
not dependent so much upon the rate of decay as upon the basic change in combustion
conditions within the furnace itself.
In Test 33-2 the furnace temperature was intentionally held constant. To achieve
this, it was necessary to cut down on the underfire air to the furnace. In this case also,
the poor test results resulted from significant changes in the combustion conditions
within the furnace rather than from the effects of turbulent mixing.
Finally, reductions in combustible emissions in Zones 2 and 3, of the breech were
rarely seen. We propose two reasons for this phenomenon:
The gases in Zones 2 and 3 emanate from portions of the furnace not
covered by overfire mixing jets, and
85
-------
Table 11. RELATIVE AGREEMENT BETWEEN
ANDv^)calc.
expt
0.284
0.150
0.100
0.25
-35.8
-5.5
+7.6
0.15
-19.7
+2.8
+11.2
0.05
-0.7
+10.1
+15.3
0.3
0.2
x
0.1
0.05
0.10
0.15
0.20
0.25
Figure 27 (1-x) vs. (0t)b.
86
-------
The temperatures in these zones are lower than in Zone 1 to the extent that
a significant portion of the concentration distribution could extend into the
quench zone. Unfortunately, the installation of overfire mixing jets in these
colder portions of the furnace would not be effective since the positive
effect of the increase in mixing intensity would be offset by the negative
effects of extending an even greater portion of the concentration distribu-
tion into the quench zone.
RESULTS OF PARTICULATE TESTS
As a complementary portion of the test program, pollutant emissions were measured in
the incinerator stack for particulates, HC1, SOX, and NOX. The particulates were analyzed
for trace metal content and polynuclear aromatic hydrocarbons.
The emissions were measured using the standard particulate sampling techniques (also
called Method 5) as described in the Federal Register of December 23, 1971. Measurements
were made during the time of the first five BASE tests. The gaseous species (HC1, SOX,
NOX) were measured in the duct leading to the base of the stack using standard impinger
collection techniques. Particulates were collected at a point in the stack 60 ft. (18m) above
the foundation of the stack. Pertinent stack data at the sampling point follow:
stack O.D. 20 ft (6.1m)
stack I.D. 12 ft (3.6m)
diameters upstream 10
diameters downstream 3.4
Because of the location of the existing sampling ports, measurements were made by
traversing the stack on one diameter from both sides; a total of 40 traverse points were used.
The location of these points in each radius is given in Table 12.
Results of the gaseous pollutant measurements are given in Table 13. The data are all
normalized to 12% CO2. The values are expected to vary with refuse composition as
observed, but there is a very wide range in the HC1 values from 220 ppm.
The basic stack particulate and sampling data are given in Table 14. The particulate
data are summarized in Table 15 on the basis of 12% CO2. The amount of particulate below
the 5- to 10-jz cyclone cutoff and collected on the filter is 83%, 80%, 84%, and 78% in test
samples 2-5, respectively.
87
-------
Table 12. SAMPLING POINTS PER RADIUS
Point Number Distance from Inner Wall
1 5ft 0-5/8 in
2 4' 4-3/8"
3 3' 10-1/8"
4 3' 5-7/8"
5 3' 1-7/8"
6 2' 10-3/8"
7 2' 7"
8 2' 4"
9 2' 1-1/4"
10 1' 10"
11 V 8"
12 1' 5-3/8"
13 1' 4-1/8"
14 1' 0-7/8"
15 10-5/8"
16 8-5/8"
17 6-5/8"
18 4-1/2"
19 2-3/4"
20 7/8"
Table 13. GASEOUS POLLUTANT EMISSIONS
(at12%CO2)
Concentration (ppm) for Test
Species 1 234
NOX 130 85 220 125 170
SOX 1,170 530 240 970 38
HCI 28 220 81 8 2
88
-------
Table 14. STACK PARTICULATE DATA
Test
Volume Samples (m3)
Stack Velocity (ft/sec)
Stack Temperature (°F)
Stack Static Pressure (in H20)
Moisture (%)
C02(%)
Particulate Weights (g)
Probe + Cyclone
Filter
Aqueous Residue3
Organic Extract8
a. of impingers
2
4.63
25
389
0.02
21
2.7
0.127
0.634
0.161
0.112
3
2.94
.14
450
0.02
33
4.1
0.178
0.172
0.108
0.103
4
2.12
15
380
0.02
28
3.1
0.044
0.225
0.070
0.018
5
3.10
14
470
0.01
30
5.0
0.160
0.580
0.045
0.032
Portion
Table 15. PARTICULATE ANALYSES
(at12%C02)
Particulate Weight (g) for Test
Probe + Cyclone
Filter'
Aqueous Residue
Organic Extract
Total Probe + Cyclone
+ Filter
0.572
2.850
0.725
0.505
0.516
2.060
0.303
0.300
0.171
0.878
0.273
0.072
0.384
1.390
0.108
0.077
3.422
2.576
1.049
Particulate Concentration based on Total of Probe + Cyclone + Filter
g/m3 (uncorrected) 0.16 0.30 0.13
g/m3 at12%C02 ,
0.71
0.87
0.51
1.774
0.24
0.58
89
-------
Analysis for trace metals in each of the filter and probe plus cyclone fractions was
carried out by spark emission spectroscopy. These results are given in Table 16 for the filter
portion and Table 17 for the probe plus cyclone portion. There is a fair degree of similarity
between the samples of each type, but some large differences in the relative compositions of
the filter and probe plus cyclone catches.
Table 16. TRACE METALS ANALYSIS OF FILTER CATCH
(values in ppm of participate)
Base
Element
Al
Ag
B
Ba
Ca
Cd
Cu
Cr
Fe
Ga
K
Mg
Mn
Mo
Ni
Pb
Sn
Si
Sr
Ti
V
Zn
Na
2
22,000
30
560
5,600
250
140
30
80
10
42,000
800
300
20
3
17,000
1,700
560
30
100
5
25,000
42,000
3
39,000
50
3,500
7,800
400
200
80
200
20
58,000
1,500
400
60
10
23,000
2,700
1,200
80
200
10
35,000
58,000
4
40
2,300
9,100
200
200
20
100
10
11,000
2,300
500
30
23,000
5,500
3,600
30
82,000
11,000
5
52,000
30
2,300
2,900
1,700
170
10
100
10
20,000
3,500
900
10
20
20,000
4,000
2,900
20
69,000
8,600
90
-------
Table 17. TRACE METALS ANALYSIS OF PROBE AND CYCLONE CATCH
(values in ppm of paniculate)
Element
Base
At
Ag
B
Ba
Ca
Cd
Cu
Cr
Fe
K
Mg
Mn
Na
Ni
P
Pb
Sb
Sn
Si
Sr
Ti
Zn
100,000
30
30
1,000
100,000
30
300
300
100,000
3,000
10,000
1,000
3,000
100
3,000
1,000
100
30,000
30
3,000
10,000
100,000
100
30
1,000
100,000
100
1,000
300
100,000
10,000
10,000
1,000
3,000
100
3,000
30,000
1,000
1,000
30,000
100
10,000
30,000
100,000
100
30
1,000
100,000
100
3,000
300
100,000
10,000
10,000
1,000
3,000
100
3,000
10,000
300
1,000
30,000
100
10,000
30,000
30,000
100
1,000
30,000
100
300
300
100,000
3,000
3,000
300
3,000
100
1,000
10,000
300
300
10,000
30
10,000
30,000
91
-------
SECTION VIM
DESIGN GUIDELINES
Up to now the report has been focussed on the description of the combustion process
using various mathematical models to delineate the fluid flow and patterns within the
furnace, the effect of different types of mixing, and the importance of air distribution and
temperature control. All of these major topics can be focussed directly on the design
guidelines for incinerators aimed at reducing combustible emissions and improving the
combustion efficiency. The important parameters are furnace temperature, furnace design,
and the air distribution and flow rate. We will concentrate first on these parameters and
then specifically on the criteria for the design of an overfire jet mixing system to take
advantage of the several benefits such a system offers.
TEMPERATURE
The plot of emission factor as a function of temperature, which was given in Figure 1,
identified two phenomena which could cause significant levels of combustible emissions:
quenching, and
oxygen deficiency resulting from a mixing limitation.
Quenching became prominent at temperatures below 1200°F, while the limitations in high
temperature mixing were observed at temperatures above 1800°F. There was apparently no
difference in the combustible emission levels in the range from 1400°F to 1800°F. As a
basic operating guideline, the breeching should therefore be maintained within this tem-
perature range at all times.
, . , Table 18, EXCESS AIR REQUIREMENTS FOR
These temperatures correspond to a EFFICIENT COMBUSTION
range in CO2 concentration of approximately
6 to 10% and a range in excess air between Recommended
100 and 200%. The latter figure can be placed Fuel Excess Air (%)
in perspective by considering the recom- Natural Gas 5- 10
mended percent excess air requirements for Fuel Oil 10- 20
efficient combustion of other fuels as shown Coal 30' 60
in Table 22. Refuse 10°-200
FURNACE DESIGN
The strategy for combustible emissions is dependent upon maintaining control of the
various operating parameters, relying on design parameters such as furnace configuration
and relative dimensions. The designer can considerably enhance the control and mixing
93
-------
required to reduce emissions. The furnace should be air-tight to avoid in-leakage that can
quench the combustion or alternatively avoid smoke escaping from the furnace chamber. In
sealing the furnace, however, care should be exercised to ensure that the grates are still
readily accessible for maintenance.
Configuration
One of the important functions of the furnace configuration with respect to eliminat-
ing combustible emissions is to ensure that the gases are uniformly mixed throughout the
furnace to minimize stratification such as that observed at Newton. In our tests, the data
clearly showed that the hot gases from the burning zones formed a stable layer in the upper
portion of the breech, while the colder gases from the burnout zone stayed at the bottom of
the breech. This was evidenced by the existence of temperature profiles within the breech
and the existence of high H2 O concentrations in the colder gases caused by water evapora-
tion in the ash quench hopper. The stratification is indicative of the existence of large
temperature gradients in the furnace and suggests that zones within the furnace exist where
quenching does occur. The test data clearly showed that there was a quench zone over the
burnout grate and indeed high levels of combustible emissions were measured. To avoid this,
the configuration of the furnace should be such as to induce front to back mixing, forcing
the hotter gases to pass through the colder gases or vice versa. Several designers have
suggested furnace configurations specifically aimed toward these ends.
Two configuration attempting to achieve this type of mixing are shown below in
comparison with the Newton configuration.
Hot gases forced through cold gases
Height
Cold gases forced through
hot gases
Hot and cold gases remain
stratified
The height of a furnace should be as close to the refuse surface as possible without
effecting flame impingement on the roof. There are three reasons for this constraint:
1. The buoyancy of the flame causes the gases to accelerate at a rate which is
dependent upon the height. The lower roof will effect a lower gas velocity,
permitting the overfire air stream to influence the overall mixing patterns
more strongly. This would result in better mixing control rather than
94
-------
allowing gases from the bed to dominate the flow patterns in the furnace
completely.
2. The lower velocity makes it easier to achieve a full jet penetration of the
furnace without introducing so much air that the flame would be quenched.
This is important in adding flexibility to the jet design.
3. The lower gas velocities result in smaller negative pressures at the base of the
flame and therefore minimize leakage.
Furthermore, we believe that the height of the roof above the refuse bed should remain
approximately constant throughout the length of the furnace. On many furnaces the roof is
at a constant height, but each grate is successively lower than the one preceding it so that
there is a great difference in the height above the grate at the front and the back of the
furnace. This results in the possibility of establishing pressure gradients within the furnace.
We observed in-leakage over the third grate at times when there was out-leakage over the
first grate which confirms this concern.
Length
The length of the furnace is one of the primary parameters used in setting the refuse
residence time. To achieve a good burnout, we believe that residence times on the order of 1
hour are required. Shorter residence times could be used if burning rates higher than 60
lb/hr/ft2 were attained. In many configurations the length of the furnace also is the
predominant factor in determining the gas residence time. The required value for the
gaseous residence time is discussed below in conjunction with the jet design; times on the
order of 1 second are not unreasonable in large incinerators. Smaller units having much
higher mixing intensities have demonstrated effective residence times as low as 0.5 second.
Width
The width of the furnace is of primary importance in specifying jet penetration
discussed below, but it is more often set to attain a desired feed rate for a given furnace
length and burning rate.
UNDERFIRE AIR DISTRIBUTION AND FLOW RATE
The primary air flow in the furnace is the underfire air. The capacity of the underfire
air system should be at least 150% of stoichiometric air requirements with the flexibility of
being able to deliver 100% of stoichiometric requirements under any given section of the
grate, especially in the sections of the grate where most of the burning occurs.
Each section of the furnace should have two important design features. First, the air
plenum should have the capability to subdivide and individually control the air rates and,
95
-------
second, the grates should be designed to allow an even distribution of air over the entire
surface of the grate. The latter requires that the wind box be sealed so that high undergrate
pressures can be attained, and that the grate have a high pressure drop with respect to
refuse. In Newton, the grates have a relatively low pressure drop and the non-uniform
distribution of underfire air was clearly seen. In particular, we observed the channelling of
air around rather than through the bed.
The theoretical air requirements for various fuels, including refuse, are given in
Table 23. As a rule of thumb, 10 scf of air are required per 1000 Btu's of heat liberated. The
total air requirement is calculated from the equation :
Q . (HHV, « s,cic, (TPD) - (50,
1 000 Btu 1 440 mm/day
where
Q = fan output
HHV = refuse heating value (approximately 4500 Btu/lb).
For example, a 250-TPD furnace requires 23,400 scfm at 150% of stoichiometry (50%
excess air). The air handling capacity of the grate is set by the burning rate rather than the
feed rate:
Table 19. THEORETICAL AIR REQUIREMENTS OF REFUSE AND OTHER FUELS
Fuel Ib Air/lb Fuel Ib Air/MM Btu scf Air/M Btu
Refuse 3.22 724 9.77
Wood 3.29 703 9.49
Peat 2.33 722 9.75
Lignite 5.27 746 10.07
Sub-Bituminous B 7.58 740 10.00
Bituminous (High Volatile) 9.08 742 10.01
Bituminous (Volatile) 10.99 760 10.26
Anthracite 9.23 831 11.22
Fuel Oil 13.69 750 10.13
Methane 17.26 722 9.75
96
-------
For a burning rate of 60 lb/hr-ft2, the underfire rate is 68 scfm/ft2 or about 8000
scfm under a 8- by 15-foot grate. Burning rates much higher than this are possible and the
designers must ensure that the grate is not limited by gas handling capacity. A safety factor
of 5 applied to the burning rate, i.e., assuming a rate of 3000 lb/hr-ft2 , is not unreasonable
in the absence of actual performance data for a particular design. The impact of this design
strategy is felt in these areas: burning rates, temperature control, and mixing.
Burning Rates
Our test data showed that the burning rate tended to remain constant at approximately
60 lb/hr-ft2 , the reason being that the bed requires a minimum amount of oxygen to burn
and naturally will draw in air from overfire to make up oxygen not supplied with the
underfire air. As the underfire air rate increased, the overfire air entrainment decreased. We
concluded that the 60 lb/hr-ft2 was a minimum burning rate resulting from the fire's natural
tendency to draw in oxygen to sustain combustion. The relatively constant burning rate
exhibited for less than stoichiometric air flows confirms the existing design rule for setting
burning rates at 60 lb/hr-ft2, but it does not suggest that this burning rate is fixed regardless
of operating conditions. With positive control and adequate underfire air fan capacity,
reliance on the overfire air is minimized and the burning rate is no longer limited by
naturally available overfire oxygen. As a result higher burning rates are attainable. Not
enough data are available to predict limiting burning rates, but values as high as 300
lb/hr-ft2 have been reported.
Temperature Control
Underfire flow rates greater than the stoichiometric requirement also enhance the
capability to control furnace temperature, particularly within the bed and at the grate.
Consider the schematic diagram shown below representing the refuse bed temperature as a
function of stoichiometric air flow rate. At air flows less than the stoichiometric require-
ments all of the oxygen will be consumed, resulting in the generation of combustion
products at the adiabatic flame temperature. The gases leave the bed at approximately the
Refuse
Bed
Temp
100%
Percent Stoichiometry
97
-------
flame temperature less any cooling of the gases as they pass through the refuse bed. The
occurrence of overfire entrainment compounds the problem by tending to stabilize the heat
release rate, regardless of the underfire air setting.
The small slope of the curve above indicates that low underfire air rates provide very little
refuse bed temperature control. On the other hand, for flow rates greater than the
stoichiometric requirement, the additional air tends to dilute the combustion products so
that an increase in the air flow results in a fairly rapid decrease in the bed temperature. This
type of temperature control would be of considerable importance to the incinerator
operator, but the concept is not currently in practice except in the newest units which
provide adequate underfire air. Care must be exercised, however, since a large amount of
excess air promotes quenching and particulate entrainment, particularly over the burnout
section of the grate. The latter concern is minimized because most incinerators will soon
have high efficiency air pollution control devices, and also because oxygen provided by
overfire air currents tends to entrain particulate.
Mixing
Control of the underfire flow rate permits much better temperature control over the
bed and effectively eliminates large variations in temperature. As a result quenching zones
are eliminated and stratification is minimized.
OVERFIRE AIR JETS
The advantages of using overfire air jets are temperature control, increased mixing, and
possibly slag reduction. During the course of this work it was very apparent that the sidewall
jets were much more effective in controlling temperatures in the 1400°F to 1800°F range
than were the existing roof jets, the reason being that jets, when properly designed, attain a
full penetration of the furnace gases and result in a much more uniform mixing than do the
roof jets. In addition, the jets can be effective in inducing a uniform side-to-side mixing,
which is easily observed as well as increasing the mixing intensity. However, because of the
furnace dimensions we believe that front-to-back mixing will be more readily achieved by
proper furnace design. During the tests we were not able to demonstrate a significant
increase in mixing intensity; however, from these tests we have derived criteria for evalu-
ating the types of furnace conditions under which overfire air jets could be effective in
inducing the fine-scale turbulence required to improve the combustible emissions burnout.
Additional Uses for Jets
Finally, others have suggested that jets can also be effective for reducing slag formation
on incinerator walls. We were unable to confirm this observation. Certainly the early
tempering of the gases with side wall jets rather than roof jets favors reduced slag formation,
but the increased momentum of the particles induced by the high jet velocities can easily
98
-------
increase the collection efficiency of slag on impact with the wall. Hence, care should be
exercised in placing the jets to avoid inducing impingement of entrained particulate on
furnace walls.
We did observe an increased rate of slag buildup directly over the jet opening and
recommend the jets be surrounded by a slag-resistant refractory, such as silicon carbide.
Design Criteria
The two parameters of importance in designing overfire air jets are the jet penetration
and the position of the jets. The method of Ivanov for determining jet penetration was
verified as a result of the testing program. With respect to the jet configuration or spacing,
several variations were tried including both opposed jets, interlaced jets, and jets acting only
on one side of the furnace; all were equally effective. An additional point, not considered in
earlier studies is that the jets should be located as close as possible to the burning surface for
the following reasons:
a. During the tests one could clearly observe a reduction in the flame height as
a result of the overfire air jets. In furnaces designed with relatively low roof
heights, as we have recommended, this aspect is very important for limiting
the amount of flame impingement on the roof, thereby avoiding potential
refractory damage.
b. The temperature control and uniform mixing are more easily achieved when
the gas velocity off the refuse bed is relatively low. This permits a greater
control of the macroscale mixing and also results in the attainment of a full
penetration with a minimum of overfire air so that quenching can be
avoided.
The procedure for designing jets is outlined as follows:
a. Determine overfire air requirement based upon temperature control -
Example:
Furnace size - 250 TPD
Stoichiometric air 15,600 scfm
Underfire air (approximately 150%) - 23,400 scfm
Overfire air (approximately 100%) - 15,600 scfm
b. Set jet size and number -
In practice, the jet diameter is limited in size only by the tendency of the jet
to slag over. A secondary effect on jet power output, as discussed below, has
99
-------
a significant effect only on small-diameter jets, but the design is usually
constrained by the slagging problem. A 3-inch jet diameter is about the
smallest practicable. Jet spacing should be between 3 and 5 jet diameters.
Example:
Furnace length 40 feet
Jet diameter 4 inches
Spacing 16 inches
Number = 40 feet/1.33 feet/jet = 30 jets
c. Set jet penetration and configuration
The various possible configurations jets on one side, jets opposed on both
sides, jet interlaced on both sides appeared to be equally effective. Ivanov
has suggested that jets on only one side have a slight advantage. The selection
of a configuration will depend upon considerations such as slag reduction.
Example:
Configuration one side
Penetration (furnace width) - 8 feet
d. Determine jet velocity
Since the total overfire air and the number of jets are specified above, the jet
velocity can be calculated. Using Eq. (6), the required crossflow velocity can
be calculated as a function of penetration:
ui = Qr/N|Lj J _ 15,60Qscfm = 100 ft/sec.
V4 / 7r(30)(60)/l/3fP\ (52)
e. Calculate required crossflow velocity (Eq. (4)):
2260 R\_ 11 -7 ni (53)
^
100
-------
f. Estimate jet height above refuse (Eq. [9]):
Temperature ratio* - 2260°R/540°R = 4.2
23,400
uc = Q/Ag = (40> x 8')60 = l -2 ft/sec-
1ii _ i (54)
TC J
L
= [(13.7)2 -(1.2)2 ft2/sec21/2.32 ft2/sec2 (3.2) = 0.9 ft
The outline above illustrates only one method for arriving at a consistent jet design.
Based upon an initial air flow criterion, the order of the calculations could be changed to
base the design procedure on a mixing criterion, considering the primary jet parameter to be
jet penetration rather than total overfire air rate. An example in which this design approach
would be more useful would be designing the mixing system for controlled-air incinerators
where the underfire air is severely limited so that the introduction of large amounts of
overfire air does not result in quenching.
Mixing Intensity Criteria
One of the methods for reducing combustible emissions is through an increase in the
intensity of turbulence in the overfire region. The power required to attain a given level of
mixing depends upon the relative densities of the fluids being mixed, but more importantly
upon the size of the furnace. As a general rule, the required mixing power scales with the
fifth power of the characteristic dimension of the system. Equation (24) gave the
relationship for 0 in terms of air flow rates.
The total jet flow rate is determined as a product of the individual jet flow rate
(Eq. (44)) times the grate area. In most cases the grate area is simply the product Lp Lf.
Substituting in these values the total flow rate QT : is determined from Equation (55):
(55)
The above relationship is dependent upon such things as the spacing parameter and the
furnace dimensions, but is not dependent upon the jet diameter or jet velocity. Hence, in a
properly designed jet system the total air flow rate introduced via overfire air jets depends
*Note: In a sealed furnace the temperature ratio becomes 2260°R/1460°R or 1 .55 so that:
u = 8 ft/sec.
U
Az = [ (82 - 1 .22 )ft2 /sec2 ] /2 (32 ft/sec2 )0.55 = 1 .8 ft
101
-------
upon the furnace and combustion parameters, but not on the jet design itself; i.e., the total
flow rate of the jets is fixed for a given application. The power output of the mixing jets,
however, is a function of the jet design, specifically being inversely proportional to the cube
root of the jet diameter. For example, if a jet system is designed using 1-inch jets rather than
4-inch jets, the power output is increased by a factor of approximately 1.6. In most cases,
the jet design will be constrained more by the occurrence of slagging preventing the use of
small-diameter jets than by the microscale power input considerations. For this reason and
others the microscale considerations do not play an important role in the jet design itself.
However, the microscale relations can be of considerable importance in determining which
furnace designs could benefit from the use of overfire air jets for microscale mixing and
which furnaces would benefit from overfire jets only on the macroscale; i.e., for increased
temperature control, slag reduction, and reduced flame heights. To be effective on the
microscale the factor /3t should have a value of approximately 1/2. The procedure for
determining the potential effectiveness of overfire jets is as follows:
a. Determination of j301
The initial value for the mixing intensity is dependent upon the furnace
configuration and dimensions based upon an application of Eq. (36). The
example considered below illustrates the use of these equations.
Example:
Consider the furnace jet system in the example given under design criterion
above:
2/3
(56)
P0t = 0.15
(2700 ft3/sec) (1 sec)
(320 ft2) 8 ft
2/3
= 0.19
(57)
b. Determine mixing intensity with jets:
320 ft2 \ / 270 ft3 /sec
2.6ft2/ \2700 ft3 /sec
)'I
(58)
In this case, the incinerator would benefit from an installation of jets.
102
-------
REFERENCES
1. Systems Study of Air Pollution from Municipal Incinerators, Arthur D. Little, Inc.,
Report to EPA, Contract CPA-22-69-23, March 1970.
2. Incinerator Overfire Mixing Study, Arthur D. Little, Inc., Report to EPA, Contract
EHSD 71-6, February 1972.
3. Ivanov, Yu V., Effective Combustion of Overfire Fuel Gases in Furnaces (translation),
Published by the Estonian State Publishing House, Tallinn, Estonia, U.S.S.R., 1959.
4. Hawthorne, W. R., Weddell, D.S., and Hottel, H.C., Mixing and Combustion in
Turbulent Gas Jets, Proc. of Third Symposium on Combustion and Flame Explosion
Phenomena, the Wilkins and Wilkins Company, Baltimore, Md., 1949.
5. Corrsin, S., Simple Theory of an Idealized Turbulent Mixer, Jour. A.I.Ch.E., Vol. 3,
No. 329, 1957.
6. Turbulent Flows and Heat Transfer, Volume V: High-Speed Aerodynamics and Jet
Propulsion, C.C. Lin, Ed., Princeton Univ. Press, Princeton, N.J., 1959.
7. Turbulence Classic Papers on Statistical Theory, Friedlander, S.K., and Topper, L.,
Eds., Interscience Publishers, Inc., New York, 1961.
8. Comprehensive Test Program Particulate Emissions and Baffle Washer Performance
at the Lower Merion, Pennsylvania, Incinerator and the Newton, Massachusetts, Incin-
erator, Report for the Research Corporation of New England, Metcalf and Eddy, Inc.,
Boston, Mass., 1970.
102a
-------
APPENDIX A
TEST MEASUREMENT AVERAGES
Within this appendix, the averages of all the test measurements made during the testing
program have been tabulated. The analytical data are listed in Table A-l. A summary of the
various operating conditions under which each test was run is given in Table A-2 for the
Series 20 tests, and in Table A-3 for the Series 30 tests. Finally a summary of the material
and energy balance is given in Table A-4.
In addition, we have included actual test data from the initial four-week series of tests
during which time the analytical procedures were being verified. These data are summarized
in Table A-5. During these tests we noted severe probe leakage problems and difficulties in
measuring velocity data. Therefore, the absolute quantities listed in the table are often
incorrect, as can be noted by comparing the reported value of CO2 (corrected for moisture)
with the data given in Figure 26. However, the relative values should still be accurate and
can be used as additional data for verifying such things as the existence of the water gas shift
equilibria. Abnormally low velocities (less than ~ 35ft/sec) are a result of velocity probe
plugging. Data in Figure 28 can be used to obtain converted readings. Finally, the measured
H2O was usually lower than the actual moisture content during the first weeks of testing
before the new analytical technique for H2O was used. Data in Figure 25 can be used to
obtain corrected readings for the value. We have not relied on these data for confirming the
various models presented, but we have included them in this report because much of the
data is valid; considering the tremendous lack of actual test data on a municipal incinerator,
even inconsistent data (when recognized as such) convey more information than has been
available previously.
103
-------
Table A-1. ANALYTICAL DATA
DATA
BREECH ZONE 1
Run
BL
JET
BL
BL
JET
BL
JET
BL
JET
JET
BL
BL
JET
JET
BL
- 22/1
- 22/1
- 22/2
- 23/1
- 23/1
- 23/2
- 23/2
- 24/1
- 24/1
- 24/2
- 24/2
- 25/1
- 25/1
- 25/2
- 25/2
CO2
10.77
8.73
10.43
9.76
7.25
8.90
5.70
6.61
9.39
8.71
10.03
11.80
8.35
7.74
8.45
% (Wet)
02
6.07
8.95
7.08
8.36
11.02
7.85
11.33
10.44
7.73
9.14
6.53
5.40
7.11
11.80
10.78
PPM (Wet)
H2O
14.50
12.20
15.20
13.80
10.04
16.20
9.50
14.40
14.00
10.80
16.20
12.90
14.30
7.80
11.90
CH4
5.13
3.29
11.87
12.28
1.19
1.46
0.90
2.56
3.44
0.89
1.25
22.21
26.99
1.38
1.05
CO
0.85
0.21
0.84
103.44
52.64
245.11
119.91
79.18
47.30
40.14
341.48
641.27
679.17
64.54
64.31
CO 2
5.38
5.56
7.00
5.46
4.74
3.84
4.75
6.11
6.09
5.79
5.77
8.15
6.74
5.81
3.99
BREECH ZONE 2
02
13.54
11.83
13.00
14.36
14.69
14.32
13.14
11.83
11.46
12.19
11.93
11.11
13.34
11.19
17.14
H2O
7.10
6.80
7.00
7.30
5.50
7.00
5.80
2.40
9.50
7.30
8.40
11.40
9.50
5.90
4.90
CH4
9.29
12.11
12.00
10.66
9.92
10.23
16.95
12.36
15.15
9.73
15.57
14.39
8.37
5.17
7.84
CO
455.21
270.28
142.50
308.22
359.10
1,185.75
1,071.'52
272.51
9.05
11.58
11.45
338.89
259.43
275.24
461.23
CO2
0.71
2.21
1.54
0.25
0.12
0.60
0.77
0.26
1.83
1.06
1.44
2.50
3.28
1.79
1.17
i cmrcnMi unc i ri
BREECH ZONE 3
02
18.82
18.03
18.72
19.65
20.06
19.94
19.19
18.55
16.35
17.38
16.94
17.88
17.76
19.33
19.95
H2O
3.10
3.80
3.40
4.10
2.10
2.70
2.80
4.50
5.10
3.40
3.70
5.50
5.40
3.20
2.20
CH4
4.84
6.73
14.94
2.87
3.18
10.37
20.16
6.44
23.96
21.01
35.39
30.71
42.57
24.20
22.49
CO
77.76
50.50
119.52
9.59
9.79
608.12
510.30
23.87
249.11
217.35
300.93
571.72
339.37
307.34
501.22
Zone
1
1,721
1,508
1,700
1,715
1,437
1,717
1,421
1,647
1,694
1,551
1,730
1,900
1,796
1,400
1,523
II
1,446
1,387
1,458
1,446
1,125
1,350
1,158
1,385
1,470
1,341
1,426
1,696
1,587
1,246
1,131
III
1,029
1,004
1,100
1,046
737
871
856
892
1,152
949
1,071
1,283
1,225
875
742
Avg.
1,366
1,272
1,411
1,372
1,058
1,265
1,123
1,271
1,424
1,258
1,382
1,598
1,505
1,143
1,097
BREECH FLOW
Velocity (ft/sec)
1
51.80
50.41
54.17*
53.02*
44.51*
52.12*
48.19*
55.74
56.50
51.42
53.93
54.94*
51.38
47.50*
49.31
II
48.28
46.58
47.66
51.74
44.54
49.05
49.01
47.54
50.84
44.52
47.15
54.86
52.53
44.45*
44.41
III
42.78
44.77
38.87
44.23
37.08
40.81
39.81
42.26
45.59
40.23
44.08
48.24
48.57
42.22
36.30
Total
Ib. moles/hr
7,643
8,007
7,376
7,935
8,102
8,026
8,439
8,210
7,929
7,742
7,705
7,496
7,587
8,186
8,142
BL
JET
JET
BL
JET
JET
BL
JET
JET
BL
JET
JET
JET
- 31/1
- 31/1
- 31/2
- 32/1
- 32/1
- 32/2
- 33/1
- 33/1
- 33/2
- 34/1
- 34/1
- 34/2
- 34/3
11.11
7.96
8.22
9.27
9.60
8.05
10.54
8.82
9.08
9.30
8.63
8.00
9.20
6.16
9.66
9.28
6.82
6.98
8.05
6.25
8.38
6.45
7.02
8.37
9.82
7.17
13.20
10.50
11.60
14.65
12.70
13.40
14.30
12.65
15.10
12.20
12.75
9.00
8.00
40.10
2.50
0.88
9.96
2.18
4.15
21.00
8.03
267.43
46.09
6.63
3.18
3.68
145.56
30.43
17.68
187.77
165.87
157.61
241.67
63.32
382.05
210.72
47.98
40.95
41.40
7.33
5.09
5.85
7.76
7.91
6.59
6.56
6.76
6.93
6.88
5.85
5.04
6.50
10.49
13.15
12.63
10.56
8.89
9.98
12.02
11.26
11.04
10.92
11.53
12.62
11.33
9.50
7.35
7.10
9.70
11.10
10.25
7.50
9.85
8.70
8.20*
8.45
6.50
7.10
28.96
4.16
25.54
6.32
5.33
9.87
10.82
8.11
219.12
5.69
11.62
9.81
12.07
162.90
187.61
278.70
31.60
120.01
229.49
300.62
234.39
314.98
183.60
244.89
250.11
222.96
2.65
1.25
1.61
2.26
3.30
2.31
2.06
3.09
2.98
1.63
1.41
1.90
1.34
16.38
18.10
17.99
14.90
13.78
15.33
17^00
16.21
16.12
16.09
15.97
16.40
16.86
5.30
3.20
4.80
5.65
8.10
7.60
6.05
6.25
6.80
3.60
5.50
4.60
4.20
38.44
5.42
4.76
32.07
34.46
62.64
42.93
76.87
950.64
17.83
28.06
73.93
21.07
181.82
20.32
19.04
264.18
22.97
200.60
420.42
517.96
829.48
337.40
304.76
453.15
411.94
1,750
1,507
1,525
1,825
1,683
1,698
1,777
1,666
1,753
1,793
1,554
1,497
1,597
1,543
1,261
1,308
1,707
1,667
1,609
1,545
1,485
1,555
1,478
1,417
1,367
1,481
1,193
836
1,025
1,381
1,408
1,344
1,251
1,240
1,306
1,150
1,106
1,023
1,115
,477
,179
,267
,621
,576
,538
,503
,446
,512
,458
,354
,277
,398
53.26
52.28
44.20
54.55*
51.33
51.10
48.02
45.77
45.45*
53.93*
49.85*
48.85*
49.93
48.91
47.91
42.85
51.22
52.42
50.28
46.25
43.73
43.00
47.43
42.43
44.86
48.65
43.93*
38.24
38.25
48.64
48.76
46.95
42.79
42.78
46.69*
41.61
37.19
41.10
35.52
7,369
8,231
7,086
7,258
7,324
7,260
6,824
6,783
6,723
7,280
6,984
7,584
7,023
'Approximate-based on estimates due to plugging of velocity measurement probe during a run.
104
-------
Table A-2. SUMMARY OF VARIOUS OPERATING CONDITIONS - SERIES 20 TESTS
Operating Parameters
Analysis of Operation
BL
Jet
BL
BL
Jet
BL
Jet
BL
Jet
Jet
BL
BL
Jet
Jet
BL
Purpose
22/1 Exploratory runs
22/1 Exploratory runs
22/2 Exploratory runs
23/1 Exploratory runs
23/1
23/2
23/2
24/1 Exploratory runs
24/1
24/2
24/2
25/1 Exploratory runs
25/1
25/2
25/2
Procedure
No sidewall or roof jets
Sidewall jets on alternative
opposed; 50% Penetration
No sidewall or roof jets same
as BL22/1
No roof or sidewall jets
Alternate, opposed sidewall jets
on-penetration ^60%
Same as BL 23/1 except smaller
UFair
Same jet condition as Jet 23/1
except more overf ire and less
UFair
No sidewall or roof
jets
Alternate opposed jets, lower
UF air from Bl 24/1
Same jet settings as Jet 24/1
Same as Jet 24/2 except no
jets on
No jets sidewall or roof
Alternate opposed sidewall jets
or penetration "u 25%
Stone feed rate and raise
to the jet and UF in Penetration
40%
Same as Jet 25/2 except no jets
Under Fire Air
CFM | (%) || (%) in (%(
8400 11 30 59
7400 10 32 58
7400 10 32 58
9800 18 31 51
9600 17 31 52
9300 17 32 51
9100 16 35 49
8400 14 37 49
7000 11 39 50
7250 10 41 49
7250 10 41 49
8450 9 43 48
7450 10 43 47
8250 9 41 50
8250 9 41 50
Over Fire
Air
(CFM)
0
4900
SJ
0
0
4300
SJ
0
5700
SJ
0
2800
SJ
2800
SJ
0
0
2200
RJ
3600
SJ
0
Refuse
Wet/Dry Feed Rate Grate Speed (FPH)
(TPH) 1 II III
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
9.8 45 45 45
9.8 45 45 45
9.8 45 45 45
8.74 40 45 50
8.74 40 45 50
8.74 40 45 50
8.74 40 45 50
9.78 45 56 55
9.78 45 50 55
9.78 45 50 55
9.78 45 50 55
9.78 45 50 55
9.78 45 50 55
8.74 40 45 50
8.74 40 45 50
Average
Res. Time Breech Temp.
-------
Table A-3. SUMMARY OF VARIOUS OPERATING CONDITIONS - SERIES 30 TESTS
BL 31/1
Jet 31/1
Jet 31/2
Purpose
Determine effect of
blocks of jets over
drying grate & 2nd
grate
Procedure
No sidewall or roof jets
run normal conditions
Demonstrate effect of Jets 1-5 on both sides
sidewall jets over the penetration 'V 50%
drying grate and drop
zone to the second
grate
Demonstrate effect of Jets 6-10 or, both sides
sidewall jets over the penetration 'Vi 50%
2nd grate
BL 32/1 Determine effect of Open roof jets, half open
placement of jets on
roof as opposed to wall
and effect of reducing
feed, U Fair, and SJ
air by same proposition
Demonstrate effect of
roof jets
Jet 32/1 Demonstrate effect of Use sidewall jets
sidewall jets alternate opposed both sides
Jet 32/2 Demonstrate effect of Reduce feed grate speed by 10%,
sidewall jets with reduce SJ air by 10%,
lower flow, lower reduce UF air by 10%
UFair, and lower
feed rate
CFM
6400
6400
Operating Conditions
Under Fire Air Refuse
Jets Feed Rate
I (%) II (%) III (%) CFM Wet/Dry (TPH)
11
11
6750 22
39
39
6800 23 31
50
50
46
31 47
5535
SJ
Dry
Dry
6400 11 39 50 4750 Dry
SJ
1200 Dry
RJ
7200 22 29 49 4195 Dry
SJ
3790 Dry
SJ
8.74
9.78
8.74
Grate Speed (FPH)
I
40
II III
45 45
40 ' 45 45
8.74
8.74 40 45 45
9.78 45 45 40
45 45 45
40 40 45
Res. Time
(hr)
.96
.96
.96
.98
.93
1.00
Average
Breech Temp.
1460
1179
1267
1619
1576
1538
Analysis of Operation
High Furnace Low Furnace
Temperature Temperature
1850
1620
1560
1890
1750
1840
1400
1290
1385
1660
1670
1560
Average Stack
Relative Temperature
Variation ( F)
large
large
med
med
small
320
330
320
340
340
med-large 360
Burnout
Efficiency Run Times
(106Btu/hr) (hr)
98.2 1105-.-1215
80.2 1425*1535
75.1 1645*1715
100.5 1040*1150
100.9
96.8
1345*1415
1545*1655
BL 33/1
Jet 33/1
Jet 33/2
BL 34/1
Jet 34/1
Jet 34/2
Jet 34/3
Determine effect of
one sided vs opposed jets
Determine the effects
of jet placement
all on one wall or placed
on both walls
Demonstrate effect
of all jets on one
wall
Demonstrate effect
of jet on both
wallj
Baseline measurement
Use no side or roof
jets
Use all jets on North wall
of incinerator
Use alternate opposed
jets configuration
Determine the effect
of jet penetration of
the operation of the
furnace
Baseline measurement
no side jets, no roof jets.
Demonstrate effect of alternate opposed jets on
partial penetration all jets set to 25% penetration
Demonstrate effect of alternate opposed jets on all
total non-overlapping all jets set at 50% penetration
penetration
Demonstrate effect of alternate opposed jets on
maximum possible jets over 2nd grate on maximum,
penetration jets over drying grate & fall atea
on about 3/4 (50% penetration)
ND-No Data
7000 21.43 33.57 45.00
7850 20 32 48
Dry
6900 23.19 31.16 45.65 4260 Dry
6900 23.19 31.16 45.65 4260 Dry
Dry
7100 20 30 50 4880 Dry
8.74 40 50 45
8.74 40 50 45
9.78 45 50 45
7100 20 30 50 2466 Dry 8.74 40 50 50
5450 22 31 47 4487 Dry 8.74 40 50 50
8.74 40 50 50
I 06
.93
.93
.90
8.74 40 , 50 45 0.93
0.89
0.89
0.89
1503
1446
1522
1455
1354
1277
1398
1840
1690
1910
1890
1630
1010
ND
1640
1550
1590
1660
1470
1400
ND
small
large
med
small
med
ND
300
300
320
320
320
310
ND
87.9 1030*1140
84.5 1250*1350
93.1 1630*1700
88.6 1100*1150
81.2 1400*1500
79.7 1530»1600
80.8 1700*1720
-------
Table A-4. SUMMARY OF MATERIAL AND ENERGY BALANCE
BL
JET
BL
BL
JET
BL
JET
BL
JET
JET
BL
BL
JET
JET
BL
BL
JET
JET
BL
JET
JET
BL
JET
JET
BL
JET
JET
JET
22/1
22/1
22/2
23/1
23/1
23/2
23/2
24/1
24/1
24/2
24/2
25/1
25/1
25/2
25/2
31/1
31/1
31/2
32/1
32/1
32/2
33/1
33/1
33/2
34/1
34/1
34/2
34/3
CARBON BALANCE
Ib moles/hr
Cln
III
As Refuse
425
425
425
375
375
375
375
425
425
425
425
425
425
375
375
375
375
375
425
425
375
375
375
425
375
375
375
375
Gout
VXU I
AsCO2
400
425
450
375
300
325
300
325
450
375
425
525
450
400
350
500
375
350
450
500
400
425
400
400
425
375
375
400
% Deviation
Cln COut
Cln
4.4
-0.8
-6.8
-1.5
21.1
12.8
19.2
21.9
-6.6
8.3
-0.5
-28.0
-6.9
-5.1
7.0
-34.0
-1.3
5.0
-7.8
-19.0
-6.9
-11.7
-9.6
3.5
-13.0
1.2
1.8
-7.6
HYDROGEN BALANCE
Hln
III
As Refuse
625
650
700
600
475
500
300
525
700
600
650
850
700
625
550
800
600
550
700
775
625
650
650
625
650
600
575
625
Ib moles/hr
Hln
III
AsH2O
525
550
575
500
375
425
475
425
575
500
550
700
575
500
450
650
500
450
575
650
525
550
525
525
550
475
475
525
Hout
V^U L
AsH20
1,200
1,150
1,250
1,250
900
1,300
950
1,450
1,500
1,050
1,400
1,450
1,400
900
950
1,350
1,100
1,100
1,400
1,550
1,500
1,250
1,250
1,200
1,150
1,250
1,000
900
% Deviation
Hln HOut
Hln
-3.6
3.5
2.8
-16.8
-5.8
-36.6
-12.5
-53.0
-15.5
2.6
- 14.9
6.5
-9.7
22.1
3.0
7.2
-2.8
-5.7
-9.4
-7.6
-30.0
-2.9
-7.7
-4.7
6.0
-17.0
5.3
21.8
OXYGEN BALANCE
Ib moles/hr
o
As Refuse
j;
250 "
260
280
240
180
200
190
200
280
240
260
330
280
240
220
310
240
220
280
310
250
260
250
250
260
230
230
250
o
AsH2O
260
270
290
250
190
210
200
210
290
250
270
350
290
260
230
330
250
230
290
320
260
270
270
260
270
240
240
260
0
As02
3,000
3,200
2,900
3,100
3,200
3,300
3,400
3,200
3,100
3,100
3,000
2,800
2,900
3,300
3,200
2,900
3,300
2,800
2,800
2,800
2,800
2,600
2,600
2,600
2,900
2,700
3,000
2,800
°0ut
vy u i.
AsC02
800
850
900
750
600
650
600
650
900
750
800
1,050
900
800
700
1,000
750
700
900
1,000
800
850
800
800
850
750
750
800
°0ut
VSUl
AsO2
2,000
2,100
1,800
2,300
2,600
2,400
2,500
2,300
1,900
2,000
1,900
1,800
2,000
2,400
2,700
1,800
2,300
1,900
1,600
1,500
1,700
1,700
1,700
1,600
1,700
1,700
2,000
1,700
°0ut
VSUv
AsH2O
600
575
625
625
450
650
475
425
475
525
700
725
700
450
475
675
550
550
700
775
750
625
625
600
575
625
500
450
% Deviation
°ln °0ut
°ln
2.1
3.2
2.3
-4.6
.8
2.0
4.9
1.8
3.0
6.4
3.7
1.5
-4.4
3.5
5.9
4.0
1.6
5.8
7.0
4.0
2.0
.6
2.2
10.2
5.6
8.0
13.3
ENERGY BALANCE
oF/% C02
260
240
240
290
290
310
320
320
260
360
350
230
260
240
260
220
260
250
260
230
280
250
240
250
260
260
270
250
[mean]
% Deviation
From Mean
[mean] - [mean]
0.9
-6.6
-9.2
11.6
10.7
20.6
21.3
23.5
-2.1
- 1.9
-2.8
-13.2
- 1.3
-8.4
-0.7
-14.3
-1.0
-2.3
1.1
- 10.4
8.0
-5.0
-7.5
-4.3
-2.0
- 1.0
1.9
-5.7
Heat
Release
In Btu/hr
8.9
86
88
92
73
87
77
90
98
81
92
104
99
76
73
98
80
75
101
101
97
88
84
93
88
81
79
80
107
-------
Table A-5. COMPILATION'OF TEST DATA
RUN MOISTURE
1
2
2a
3
4a
4b
5
6
7.
7b
8
9«
9b
10s
lOb
lOc
11
12
13*
13b
14
WET
WET
WET
DRY
WET
WET
DRY
DRY
DRY
WET
WET
WET
WET
DRY
DRY
DRY
DRY
DRY
DRY
DRY
DRY
GRATE
1
40
35
35
35
35
35
55
55
55
55
35
50
50
55
55
55
35
55
45
35
35
SPEED ,
2
30
40
40
40
35
35
55
55
55
55
30
50
50
50
50
50
35
55
45
50
50
FPH
3
30
45
45
45
35
35
55
55
55
55
30
50
50
50
50
50
35
55
45
55
55
BED 1
1300
3780
3780
3780
980
980
915
1440
785
785
965
3160
4170
2350
2090
1960
915
785
1440
330
655
2_
2640
7170
7170
717)
4550
4550
3220
5250
4110
4110
4000
6120
7900
5100
5100
4080
4535
4535
6425
1480
7445
3_
1410
3830
3830
3830
3230
3230
3380
4500
3100
3100
3200
3260
4220
5630
10,120
1970
3375
3490
6130
1410
7030
TOTAL
5350
14,780
14,780
14,780
8760
8760
7515
11,190
7995
7995
8165
12,540
16,290
13,080
17,310
8010
8825
8810
13,995
3220
15,130
T. TOTAL
1 2 "*
25 49
26 48
26 48
26 48
11 52
11 52
12 45
13 47
10 51
10 51
12 49
25 49
26 48
18 39
12 30
25 50
10 51
9 51
10 46
10 46
4 49
j
26
26
26
26
37
37
43
40
39
39
39
26
26
43
58
25
39
40
44
44
47
OVERFIRF.
ATR. TFM
^^j^..^^
1. STOICH10METRIC CO,
1 7 1 TIVTAT £.
8
25
25
25
7
(,
3C
J
3.5
JC
20
10
9
7e
. 3
2
f.
16
48
48
48
30
30
16
24
18
18
27
30
39
23
23
18
30
20
34.
10
50
9 33
26 99
26 99
26 99
22 59
22 59
15
20
14
14
22
16
21
25
45
9
23
10
5 33
10
48
35
50
35.5
35.5
55
61
80
58
77
36
59
39
75
22
102
3650
6020
6020
6020
3140
3140
3635
6580
2735
2735
4335
4850
4450
11,470
9440
4200
4215
4365
5850
2000
5850
x.
11.0
.80
.20
.25
.30
.40
.83
. 36
.45
.49
. 24
o
.14
.48
.21
o
9.52
12.0
7.8
10.7
4.9
°2
X
8.00
20.0
18.0
19.5
19.0
19.2
18.9
19.4
19.4
19.7
21.0
20.0
20.0
19.0
20.0
20.0
11.57
7.6
11.5
8.1
12.4
ZONE
H20
^
s*
15.0
3.0
3.0
3.0
3.7
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
10.6
16.4
9.9
9.9
9.4
!*?' ZONE II
CH4
1210
10
8
3.6
8.3
32
28
17
10
12
7
8
10
12
1.7
15
20
740
195
42
41
CO
PPM
1880
65
20
10.8
18.3
59
124
61
24
73
8
21
15
30
20
13
38
2975
3000
236
325
H,
z
>
1650
0
0
0
0
30
88
29
10
47
38
16
19
5
42
21
56
1555
1800
14
149
co'lgl
^3T
'Hi*
1
2.16'
3.00;,
6.13
slstf
2.fjj
(p'!
c|w
('*@
(.2T)
(0)
3.66
7. 18
8.50
5.17
5.1
02
%
10.5
20.5
18.0
19.7
19.0
19.8
17.8
16.9
14.0
14.6
15.8
(20.0)
(20.0)
(19.0)
(20.0)
(20.0)
11.6
12.6
10.7
14.4
15.5
H20
^.
22
3
3
3
3.7
3
5
5
9-
9
5
(5)
(5)
(5)
(5)
(5)
6.2
10.9
9.9
9.9
6.2
CH4
1333
12
8.3
4
8
32
42
53
58
68
32
(8)
(10)
(12)
(1.7)
(15)
24
201
142
16
16
CO
.PPM
1738
78
23
12
12
45
253
355
495
628
344
(21)
(15)
(30)
(20)
(13)
12
2750
2500
210
270
H2
-^
14
0
0
0
30
41
217
215
78
96
25
(16)
(19)
(5)
(42)
(21)
38
1250
1800
14
39
co2
4^-
7.00
2.90
1.75
3.85
1.8
2.73
3.45
2.76
4.10
2.40
.94
1.10
.22
2.40
.21
0
1.65
4.60
4.50
3.0
2.28
°2
/,
13.0
18.5
16.8
16.3
18.0
17.4
16.4
17.1
15.6
18.1
15.9
19.0
20.0
18.0
20.0
20.0
19.4
15.3
15.9
17.0
18.4
ZONE III
V
»
8
5
5
13
3.7
5
6
6
9
9
3
3
3
4
4
4
2.5
6.2
5.1
5.1
4.5
CH4
34.2
42.0
36.4
50.5
38.9
33.7
49.2
53.6
38.2
32.7
28.0
7. 4
3.8
42.7
45.1
51.5
31.4
39.7 .
38.2
25.0
22.0
TOTAL CFH ,
X 10°
5.27
9.67
9.27
11.02
9.21
7.95
7.58
6.85
9.04
7.73
7.74
2.71
2.12
6.58
6.08
6.79
7.57
6.18
8.07
3.86
3.32
NOTE
: Number» in parenthesis were estimated from general profiles.
i;D8
-------
APPENDIX B
CALCULATION OF UNMIXEDNESS FACTOR OF REFUSE BED OFFGAS
The ratios of moles of various components with respect to moles of offgas are shown in
Table B-l below:
Table B-1. MOLE RATIOS
Unreacted Mixture Reacted Mixture
Moles Percent Moles Percent
100% . 100%
""-- * -Combustion*
^^ * ;!* ^»wi i ihsut*viwi i i
1 + 4.76nQ2 + nx Products 1 + 3.76nQ2 + nx
(4.76nQ , 100% 3.76nQ 100%
Stoichiometric 4-7fin0 "2 3.76n
Air 21+4.76n02+nXs 2 1 + 3.76nO2 +
U2 As 02
n
nx 100% nx 100%
Excess Air nx Excess Air n
S 1 + 4"76no2 + "xs X$ 1 + 3'76no2 + \
Total 1+4.76nn + n 100% Total 1 + 3.76n +n 100%
°2 xs O2 Xj
Note that in a Stoichiometric mixture nx is zero. For one mole of offgas, n02 moles of
oxygen are required and therefore 4.76 n0 2 moles of air. The mole fraction of offgas in a
Stoichiometric ratio is therefore 4.76 n0 2:
moles of combustible 1
Ys~ total moles ~ l+4.76-n02
The mole fraction of offgas in a typical mixture (not Stoichiometric) is:
y = 1 . (B-2)
l-r4.76(n02+nXs)
To relate these relations to the gas analysis, consider a blending of a Stoichiometric mixture
ys Ct; the concentration of offgas in the excess air is zero. Since concentration, C, can be
calculated from the mole fraction and the ideal gas law, then:
C=y -C=y^
^s y& M y$
C=y -C,-,
PT
p (B-3)
'None of the combustion reactions involved results in a net production of mols.
109
-------
For blending,
Cs (volume of stoichiometric mixture) + O.D (volume of excess air) = C (volume of total)
Hence:
(y, ' Ct)V, = (y Ct)Vt or ysns = ynt. (B-4)
Then, for the example in Table B-l,
PV
PV
=-- = n = l+
(B-5)
Therefore:
y = ns _ 1 +4.76 'n02
7T." "»T = l+4.76(n02+nXs) (B'6)
The initial concentration or mole fraction of offgas is not always known, and it is
important to determine y/ys in terms of gas analysis after combustion. The same approach is
used as before, considering a mole of final combustion products rather than offgas. In this
case:
and
nco2
%C02|S = - -100% (B-7)
nco2
%CO2 = - - -100%
nt
Hence
<%ro. nc« 1 + 3.76 n.
(B-8)
%C02 % l+3.76-n92
%CO2|S nt 1+3.76-n02+4.76 nXj
10
-------
This can be rearranged to:
%CO2 = %CO2 ls
(1+3.76 -n02)
U+4.76(n02+nXs)-n02
%C021
(l+3.76-n02)y
Solving for y yields:
Then:
(%CO2/%C(
_l+n02(%C02,
%C02
%C02|S
M) ir i
'%C02|J |_l+3.76n02 _
r / i %C02 \n
\l+3.76n0, %C02|S/
1 ' " %co2
%CO,L n°2
When n0 is small, the equation reduces to
°2
Ay =
_%C02_\
%co2ij
Example: For the bed offgas where CO/CO2 = 0.5, n02 = 0.058, ys = 0.784
At a temperature of 1500°F, %CO2 = 7.2%.
'_J } - 2,
]+ao58. 0*3.76(OW IS-
Ay = 0.783
1 -
I./.70
15.1%
.1%
'*
Ay = 0.404
Using the simplified expression above, Ay is estimated to be 0.410.
The required oxygen, O2 req, is calculated based upon CO:
02req = n02 (CO/CO I s)
(B-9)
(B-10)
(B-ll)
(B-12)
(B-13)
(B-14)
11
-------
The available oxygen is equal to %O2 from the gas analyses.
Hence:
02 req n02 (CO/CO|g)
O2 available %02
For the example above, a CO concentration of 100 ppm results in a value of 8.6 x 10~4.
The factor >/c' /Ay is obtained from Figure 3 and is approximately 0.40.
Since
^fif = (jfF /Ay) O Ay (B-l 6)
then, the unmixedness factor for the example is:
0.40x0.404 = 0.162
112
-------
APPENDIX C
MATHEMATICAL CHARACTERIZATION OF QUENCHING
The problem of calculating the rate at which a pocket of fuel will be quenched by
reduced temperature and determining the corresponding fuel concentration at the time of
quenching may be modeled as described below:
The temperature of the gas is cooled by conduction of heat from the fuel at the
adiabatic flame temperature to the surrounding air at ambient conditions. This represents
the most rapid rate of conduction heating and other factors, such as air preheating or
radiative heating, will diminish this rate as well.
Consider a portion of fuel as shown in
the figure right in which the temperature is AIR
initially Tf. At time, t = 0 the surface is
brought to temperature T0 and held constant;
again this is the maximum gradient. The
governing boundary value problem is:
dT/at = a a2 T/a/2
T(0,z) = Tf
T(t, 0) =TO,
T(t, oo) = T0
The solution is:
Tf-T0
= l ~e rf
If this is evaluated for gases at the quench temperature, then:
Tf = 3100°F
TO = 100°F,and
Tq = 1100° F (quench temperature).
and, therefore:
erf
= 1 -(3100-1100/3100-100) Si 1/3
(C-l)
(C-2)
(C-3)
13
-------
where zq(t) is the distance from the initial surface to where the reaction is quenched. The
rate of propagation of the quenching front is:
d(zq)/dt = 0.305 v/^/T (C-4)
If one assumes that the fuel is initially at the adiabatic flame temperature (less a
reasonable amount of heat loss) and must be cooled by dilution with ambient air, then the
change in temperature and concentration will be continuous. For any combustibles to
remain following this transition, the mechanism for cooling would have to precede so that
the stoichiometric concentration Cs would not be attained before the fuel reached the
quench temperature. The above analysis shows the rate of propagation of the quench front;
the propagation of the plane where Cs is reached must be slower than 0.305 \/a7t
Otherwise, a condition would exist in which fuel and air would exist in proper proportions
and above the quench temperature; the expected result of this condition is a complete
burnout of fuel.
First, consider the case of air diffusing without reaction, the analogous case to the heat
condition problem. The governing equation for Ca is:
aca/at = Dao2ca/az2) ca +cf = cx
Ca (t, 0)= Cao,Ca(t,oo) = 0 (C-5)
Ca (0, z) = 0
The solution is:
(Ca0 - Ca/Ca0)= erf (z/ V^B (C-6)
For the case where CO/CO2 = 0.5, Cs was shown to be 0.78.
Then:
C. -C.
= 0.78 = erf (^4 Da t > (C-7)
a
C
U
in which case
zs = 0.87 >/4D,t (C-8)
114
-------
Then, when
a = 11.2 ft2 /hr (constant)
D= 9.3 ft2/hr (constant)
d(zs)
dt
= 0.79V/07T
AIR
FUEL
-z
z=o
+z
fuel + n 'air
The governing equation for each region is:
Air
= C
o2 c
-°o Tq is a certainty and
no combustible would be expected.
\
Consider next the effect of a
rapid irreversible reaction as shown in
the figure left. Since both air and fuel
cannot co-exist, there is a plane at z
where the concentration of both is
zero. For the constant molar case:
(1 + n) products
Fuel
cf + CP = CT
3 Cf/d t = Df (32 Cf/9z2) z' < 7 < oo
Cf(0,z') =CfQ
Cf (t, z') = 0, Cf (t, oo) = cffl (G9)
z = z
The solution is of the form
= aj + a2 erf
Cf =bj + b2 erf[-
(C-10)
115
-------
Solving for the boundary values, the solution is:
C. -C
l+erfU/4Dft
where
1 -
z =
(c-ii)
From the final boundary conditions, representing the reaction stoichiometry, the implicit
relation for TJ is derived:
T? Df) 1? Cf
For the case Df = Da = D, Ca = Cf =1 (pure fluids) the relation reduces to
erf
(G13)
Then for CO/CO2 = 0.5, n = 0.23 mole air/mole fuel
dt
. (Z') = 0.57 V57T
(G14)
Because of the small amount of air required, the value for T? is still not small enough for the
rate of propagation of Cs to be slower than that for T .
116
-------
The final step in the analysis is to consider the effects of a fuel temperature lower than
the theoretical adiabatic flame temperature. Such an occurrence results from two phenom-
ena. First, the combustion reactions in the bed do not go to completion so that all of the
heat of reaction is not liberated and the flame temperature is proportionately lower. The
additional amount of heat is liberated as oxygen diffuses, but the net reduction in the heat
liberated would occur if a pocket of combustibles were quenched before the reaction was
completed.
A second mechanism for having a fuel temperature lower than the adiabatic flame
temperature is a result of heat transfer from the hot combustion gases to the refuse bed
itself. This is evidenced by an increase in the temperature of the refuse bed as well as an
increase in the production of gasification or pyrolysis products. In the case of pyrolysis
which produces large amounts of hydrocarbons, the consistency with the gasification
premise for bed-burning is largely the result of the decomposition of methane to carbon and
hydrogen, which is substantially complete at temperatures above a 1000°F.
Finally, a pocket of fuel can lose heat through radiation. This mechanism tends to cool
the hottest gases and heat the colder gases and will result in a net loss of heat for the high
temperature fuel pockets.
If we assume that the temperature of the fuel is only 2000°F, as a result of the three
mechanisms above, then the left hand side of Eq. (C-3) has the value of 0.48 rather than
0.305. Hence, the speed of propagation of the quenched plane would be greatly increased;
but in this particular case, the increase would still not be sufficient to overcome the
propagation of the Cs plane. However, the assumption of pyrolysis on a larger scale suggests
that the ratio CO/CO2 could be much larger thanO.5. Hence, the factor n in Eq. (C-l 3) would
be larger and the rate of propagation for the C plane would be correspondingly slower. As
an example, consider the ratio CO/CO2 =1.0. From a material balance the mols of air
required practically double so that n = 0.5. Then the left hand side of Eq. (C-13) has a value
of 1/3. The solution for this value is identical to the initial example for the proportion of T
Eq. (C-13). Under these conditions the propagation of Cs would occur at a slower rate than
the propagation of the quench plane. When this condition is obtained, Eq. (C-9) would still be
an adequate description of the process if the boundary conditions were changed to:
Cf(t,z.)-C,
-D_
= n
-D,
ac
Z=z -
q -J
(C-l 5)
117
-------
The concentration profiles are shown schematically below:
where in this case z marks the location of the quench plane. To the right side of z the mix-
ture is above the quench temperature and, therefore, the oxygen concentration must be 0. To
the left side, the mixture is below the quench temperature so that the co-existence of both
oxygen and fuel is possible. Under these conditions the equation for Ca is unchanged but
the equation for Cf becomes:
Cf -C,
fo f
C -Q,
1 -erf
V/4D7T
1 -erf
Z- < Z < +
erf(zV4Dft)
oo < z < Z,.
(C-16)
Interestingly enough, the concentration where quenching occurs is constant as can be
deduced from Eq. (C-10)forCf by substituting zq = .305 V4a t. If for the mixture on the
right hand side of z the diffusion equation of the form for Eq. (C-5) is solved for fuel and
products individually as well as the two combined, then it can be shown that
=(cp/cf)|z=
(C-17)
The stoichiometric boundary condition in the case of quenching reflects the net consump-
tion of fuel at the boundary z . This can be solved to yield:
1 -erf
1 4-erf
Ja.
n(C -C
ao
exp
4Dat
4Df t
1 -
(C-18)
118
-------
which is similar to (Eq. 12), .but with the substitution of (Cf - Cn) for Cf and with the
lo i 'o
additional term [1 - (Cq/Cf -Cq)] representing the flux of fuel into the quenched
mixture. Equation (C-l 2) was used to solve for z . In this case, however, z is determined
from the heat conduction problem. Substituting Eq. (C-3), for example, and assuming equal
diffusivities (Da = Df = D), Eq. (C-l 8) can be used to calculate the quench concentration C .
Simplifying the above,
*(
erf (0.305 ^ ,
1 + erf (0.305 V^/D ')
C = 1C
S 2 lfo
which reduces to Eq. (C-l3) when C equals zero and Cf =Ca = 1.0. For the case when
n = 0.5, C has a value of only 0.034. If, however, z is assumed to equal 0.48^40; t, as
suggested above, then for n = 0.5, C is equal to 0.21 which is substantial.
Up to now analysis has centered around showing not necessarily that quenching will
always occur, but only that under the right set of circumstances it can occur. The
experimental data presented in Section VII clearly show that in all three zones a certain
amount of quenching is witnessed. Several additional sophistications can be added to the
analysis, including the effect of differences in diffusivities of fuel and air and the effect of
considering the diffusivities as functions of temperature rather than as mean average
constants. The latter is of particular note because diffusivities vary with T3/2, while the
conductivity is a function of T1/2. The basic conduction and diffusion equations reduce to:
Conduction Diffusion
d6 _ .. d20 ' dC _ ^ d
d di? drj2 dTj drj
Note that both the thermal diffusivity a and the molecular diffusivity D are functions of
T3'2, but because of the placement of the derivatives the differential equations are not of
the same temperature functionality and the respective solutions will reflect this difference.
Unfortunately, these equations are all highly non-linear and require a numerical solution
rather than an analytical one. Since little could be gained from this additional degree of
sophistication, the numerical analysis was not carried out.
119
-------
From the above analysis, it is clear that the quenching phenomena can cause com-
bustible emissions under the right set of circumstances so that a final refinement is required
to devise a method for calculating these emissions. In reality, the transfer between the fuel
phase and the air phase cannot be legitimately represented by semi-infinite medium since
both of these are finite. The solutions of the basic differential equations with respect to
finite boundary conditions have been worked out and will not be attempted here.
The simultaneous heat and mass transfer mechanism is not the only means by which
emissions can be produced in gases having the appropriate amount of air and sufficient
temperature to sustain combustion. In a recent paper by Sivashinsky and Gutfinger* the
hypothesis that emissions can be caused because of the finite speed of chemical kinetics was
explored. They concluded that during the course of combustion not all of the combustible
gases are consumed in the reaction zone and that a portion of the combustible component
passes through the flame front without reacting with oxygen.
Still others have suggested a stochastic model based upon eddy collisions and applying
the same type of probability equations used in the kinetic theory of gases.
The exact mechanism or combination of mechanisms is not known and research is
underway in this area. The quantification of quenching cannot proceed beyond the level
given here until a better physical description is at hand.
Sivashinsky, G. I., and Gutfinger, C., The Extinction of Spherical Diffusion Flames, Paper B5,
Proc. 1973 International Seminar on Heat Transfer from Flames, Aug., 27-31, 1973, Trogir,
Yugoslavia.
120
-------
APPENDIX D
BASIC PRINCIPLES OF COMBUSTION
Within this appendix, the fundamental principles of combustion are presented for
those unfamiliar with combustion chemistry.
Reaction Chemistry
The principles of combustion of refuse are no different than those of other fuels, and a
great deal of study has been centered around the reactions that occur (3, 4). Most
hydrocarbon fuels are parafinic and can be represented by the formula (CH2)n. Refuse is
more cellulosic than parafinic and a more accurate formula would be CHn (H2 0)m . The
equation for the complete combustion of refuse can then be given as:
H20. (D-l)
In reality the reaction does not go to completion for some fraction of their fuel and
off-gas products. Partial combustion products, such as CO, H2, hydrocarbons, and soot,
often result. The step reactions which take place either in the refuse bed or overfire region
include oxidation reactions:
C +1/202 < - > CO (D-2)
C0+l/202 < - > C02 (D-3)
H2+l/202 < - » H20 (EM)
and reduction reactions:
C +C02 < - » 2CO (D-5)
C +H20 . < - > CO + H2 (D-6)
By combining reactions [5] and [6] , one obtains:
CO + H2O < - '+ CO2+H2
Tlie last of these reactions, known as the water gas shift reaction, is of considerable
importance in describing the gaseous equilibrium in bed-burning processes.
121
-------
Reaction Kinetics
The reaction rate for a particular molecular species is dependent upon a combination
of all the above reactions. For example, the rate of reaction for carbon monoxide is given by
the equation:
(r4-rs+2r7+r8). (D-8)
.
dt
A similar equation could be derived for each of the constituents, provided that all the
occurring reactions involving the constituent were identified.
The expression for the chemical reaction rate is also dependent upon the specific
mechanism of reaction. In very simplified terms, the rate can be expressed by the law of
mass action applied to molecular reactants. For reaction [D-3] above:
r3=k3(CO)(02)'/z (D-9)
where k3 = chemical kinetic constant, and (CO), (O2) = concentration of molecular species.
Unfortunately, the mechanisms for combustion reactions are rarely as simple as the
elementary expressions above. Typically, the mechanisms involve such things as free-radical
chain reactions or other types of activated complexes. The exact mechanisms have been
worked out over years of research and will not be discussed in detail here.
All of the oxygen-consuming reactions have extremely rapid reaction rates. This has
been experimentally verified in many instances where reactions were shown to go to
completion within a matter of microseconds, and certainly the rate of reaction in an internal
combustion engine, or in a hydrocarbon explosion, attests to the rapidity of the chemical
kinetics.
The above reactions can also go in the reverse direction, and although the same
principles apply in determining the chemical kinetics for these reactions, the rate constants
are different. The ratio of the rate of the forward reaction to the rate of the backward
reaction determines the equilibrium of the reaction which is discussed in detail in a later
section.
Activation Effects
All the kinetic expressions have their bases in statistical mechanics and represent the
rate at which molecules or other reacting species will collide in such a way as to undergo a
chemical change. But molecules must not only collide before a reaction takes place, they
must also have sufficient energy to be able to rearrange chemical bonds to undergo the
122
-------
chemical change. The large majority of molecules colliding do not have this energy, and thus
the collision will not result in a chemical reaction. The process is shown schematically
below:
Energy
Reactants
Products
The average molecule is assumed to have an energy level represented by point A. Upon
collision, the molecule does not react. However, certain molecules, because of past history,
will attain energy states higher than average. When one of these reaches the energy level of
point B, then, like a roller coaster, it will seek a point of lower energy and will either return
to the normal state A, or react upon collision to form products having an energy represented
by point C. The height of the energy barrier Eais the activation energy, and the difference
between points A and C is the net energy of reaction, the latter being directly related to the
heat of combustion.
For both the forward and reverse reactions, the dependence of the reaction rate
constant was shown by Arrhenius to be exponentially dependent upon the activation energy
as shown by the equation:
(D-10)
where Ea for the forward reaction is the difference between points A and B, while Ea for
the reverse reaction is the difference between points C and B. In the case of the exothermic
reaction, such as combustion, heat is liberated and the forward-reaction activation energy is
less than the value for the reverse reaction, so that the rate of forward reaction is faster and
the net release of energy allows the reaction to be self-sustaining once it is initiated. This
suggests an additional ramification of the activation energy the concept of ignition.
Consider a combustible mixture in which none of the molecules has sufficient energy
to exceed the energy barrier for a given reaction. In this case, none of the molecules will
undergo a reactive collision and combustion will not occur. As the temperature of the
system is increased, the molecules increase in energy to the point where one of the
molecules will gain sufficient energy to exceed the energy barrier. When this happens, a
combustion reaction will occur, liberating energy, thereby increasing the temperature of the
gas, and causing additional molecules to exceed the energy barrier. The chain process repeats
again and again, and the temperature of the gas increases dramatically. The temperature at
which this series initially begins is known as the ignition temperature.
123
-------
Chemical Equilibrium
Because the chemical reactions are occurring simultaneously in both the forward and
reverse directions, the reaction will ultimately never reach the state of "complete combus-
tion." The equilibrium between the forward and the reverse reactions is determined from
the ratio of the respective rate constants.
In the case of the oxidation reaction, the heat of reaction is very high and the forward
reaction occurs at a much faster rate than does the reverse reaction. The reduction reaction,
on the other hand, has a low heat of reaction and is, in some cases, endothermic (requires
heat). As a result of the Arrhenius Equation [10] , the equilibrium constant is a function of
temperature according to the relation:
(D-ll)
This relation is shown for the various combustion reactions in Figure D-l . In the case of the
oxidation reactions, reaction [3] , for example, the equilibrium between carbon monoxide,
oxygen, and carbon dioxide at 1800°F is on the order of 107 atm"1 . For a chemical
equilibrium-controlled reaction in which the stack gas contains 12% CO2 (dry) a material
balance suggest approximately 8% oxygen. The concentration of carbon monoxide is
calculated, below:
K4 = 107atm-'/2 = (%C°2)
(% CO) (% 02 )1/2 ( 1 atm)'/2 (D- 1 2)
(% CO) = 0.04 ppm
For reaction [7] above, the equilibrium at 2000° F is equal to:
=20
7 (CO)(H20)
Diffusion Flames
Unfortunately, the overall rate of chemical reaction cannot always be controlled by the
chemical kinetics and equilibrium is often unattained. The controlling step often is the rate
at which fuel and oxygen can be brought together to react. As a good example, the rate at
which soot particles burn is controlled primarily by the rate at which oxygen diffuses to the
surface of the soot, rather than by the rate at which carbon (soot) and oxygen combine to
form carbon monoxide. Flames of this type are diffusion flames and are distinguished from
premixed flames in that oxygen must diffuse into the fuel to form a combustible mixture.
Examples of diffusion flames range from candles to bonfires. A distinction is made,
however, in the type of diffusion flame, depending upon the type of diffusion occurring.
For the small candle, the oxygen is fed in by molecular diffusion and the flame is called a
laminar diffusion flame. In the case of a bonfire, the hot combustion gases have enough
buoyancy to generate a considerable amount of turbulence and draw air into the base of the
flame by a convective transport mechanism. These flames are turbulent diffusion flames.
124
-------
o
8.8
O O
^-« r^.
r,°F
in t m
o o o o o o o
§o o o o o o
in o oo u> i- CM
ro eg ox i < i -H
-5
0.0002 0.0004 0.0006 0.0008 0.0010 0.0012 0.0014 0.0016
T' K
Figure D-1 Equilibrium constants of combustion reactions.
125
-------
In a laminar diffusion flame the combustion is controlled by molecular diffusion
described in generalized coordinates by Pick's Law:
The rate of flux toward the flame boundary is directly related to the difference in
concentrations at the boundary and the concentration at a distance from the boundary, i.e.,
the concentration gradient. The concentration profile about the flame front can be
described by a mathematically continuous function analytically determined by the simultan-
eous solution of Pick's Law and the law of conservation of mass for a given system
geometry. The solution is relatively straightforward and has been discussed at great length in
the scientific journals.
In a turbulent system the solution for diffusive transport is not so easy a problem and
the mathematical theory is still evolving around this issue. Unlike the laminar problems,
turbulent transport takes place in bulk rather than on a molecular scale, i.e., groups of
molecules tend to transfer together from one place to another via a convective transport
mechanism. The process is referred to as eddy diffusion. In a turbulent regime, a whole
range of eddy sizes exist and there are eddies within eddies. An adequate description of the
eddy characteristics requires statistical techniques of a much greater sophistication than the
analytical techniques applied to the laminar diffusion process.
The incinerator does have its own air supplies and can best be described as partially
premixed turbulent diffusion flame. However, the premixing notwithstanding, the rate of
gas phase combustion is controlled by turbulent diffusion and the principles describing
turbulent diffusion flames will apply.
Material Balances
In several studies of the nature and composition of refuse, including those specifically
related to the City of Newton's refuse, the average composition of dry ash-free refuse was
approximately CH0.4 (H2O)0.6 so that the combustion reaction is:
CH0.4 (H2O)0.6 + 1.1 (02) >CO2+0.8H2O (D-15)
The coefficients of each reactant and product indicate the relative amounts (in mols) that
are required to combine exactly during the reaction. These proportions can be converted to
actual compositions or weight of burning material using the law of conservation of mass. A
summary of the relative proportions of combustion products is shown in Table D-l. More
often the composition of a gas sample is given in volume percent which is equivalent to the
mol percent shown in the table as well. Another unit, the Orsat analysis, is usually given on
a dry basis. The interrelationship between these various quantities can be summarized by the
equations below applied for carbon dioxide balances. These equations, Ns represent the
126
-------
number of moles of gases in a stoichiometric mixture containing CO2, H2O, and the
nitrogen equivalent of oxygen required to complete combustion; i.e.:
+nH20+nN2,s
(D-16)
If nxs represents the number of moles of oxygen present in excess of that required to
complete combustion, then the total number of moles, nt, is determined from the equation:
n, = n,. + 4.76nx<,
la AS
The gas composition (is mole percent) can be calculated for CO2 as follows:
n/-./-v
C02)W =
(%C02)d =
n
'CO-
X 100%
*
X 100%
(D-18)
(D-19)
IV n
_ CO 2
's,w
X 100%
(D-20)
>CO,)
2's,d
nco-
ns~nH20
X 100%
(D-21)
Since 1.1 moles of oxygen are required per mole of carbon (Equation [D-l 5]), the
percentage of excess air is calculated by the following equation:
(%02)d
(%XS) =
n
xs
1-lnco-
X 100% = -
1.1 (%C02)d
X 100%
(D-22)
Likewise, if one assumes that the refuse initially has 0.7 inch H2O per mole of C (35%
moisture on an ash-free basis):
ns=nco2 +
+3.76 (l.
= 6.64 n
(D-23)
Finally, Equation [D-17] can be rearranged to the form:
ns 4.76 n
xs _
= 1
(D-24)
Combining Eqs. (D-23) and D-24) and rearranging the fraction so that nt and ns are
both denominators yields:
127
-------
n,
CO-
nco-
6-64nco2
4.76
= 1
(D-25)
or
15.1%
(%02)d =
21.1%
(D-26)
Table D-1. PRODUCTS OF REFUSE COMBUSTION
Fuel
Air
Refuse
C
H
0
Moisture (H20)
Total
Component
C02
H20
N2
Mols
1.0
1.6
0.6
0.7
Mols
1.00
1.50
4.14
6.64
02 N2
(required) (3.76 XO2)
1.0 3.76
0.4 1 .50
(0.3) (1.12)
1.1 4.14
Combustion Products
%Wet % Dry Weight (%)
15.1 19.5 23.5
22.6 14.5
62.3 80.5 62.0
100.0 100.0 100.0
Energy Balances
In an adiabatic process (no external heat loss or gain to the system), a stoichiometric
mixture of refuse and air reacts liberating approximately 7500 Btu/lb of dry ash-free refuse.
All of the liberated heat goes to raising the temperature of the combustion products from
T0 to the adiabatic flame temperature Ta. The energy equation is:
(nsCp )(Ta-T0)= Z (nCp)i(Ta
T0) = 7500 Btu/lb
(D-27)
128
-------
where:
HJ = mass of component i per Ib of dry ash-free refuse
ns =
n = l
n; (stoichiometric mixture)
Cp . = heat capacity of component i
The heat capacity for the stoichiometric mixture is shown in Table D-2.
Table D-2. HEAT CAPACITIES
Component
C02
H20
N2
Total
Mols
15.1
22.6
62.3
100.0
(cals/gmols '
13.0
9.0
7.4
8.61
196.3
203.4
461.0
860.7
The final temperature of the stoichiometric mixture is referred to as the theoretical
adiabatic flame temperature and is approximately 2950°F for refuse. A temperature of this
intensity can damage a refractory incinerator. The primary method for reducing the
temperature to a reasonable operating range is through the addition of excess air. The gas
temperature is lowered because the heat liberated by combustion is used to raise the
temperature of more than just the combustion products,. The energy balance is as follows:
(n, Cps) (T -T0) + (4.76 nxs C?xs) (T -T0) = 7500 Btu/lb
Equating Equations D-27 and D-28 yields:
"sCPs
T-T
*a 1o
n.C +4.76nv«C
'xs
Equation [D-29] can be simplified to:
T-T«\ 4.76 n
T-T
A l
xs
ns
~Pxs
:Ps
AH/(T-T0)
T-T
1 io
L T, ~ Tn
(D-28)
(D-29)
(D-30)
129
-------
Substituting Equation [D-26] and rating that nxs/ns equals (%O2)/100%> the equation
becomes:
T-T0\ (%CQ2)W , (%02) p.jpx. T-T0\
Ta-T0) (%C02)S>W (21.1%) Cp \Ta-Tj
The second term on the right is a correction factor for the changes in the heat capacity in
going from reactants to products and is on the order of 0.1. Since both the stoichiometric
CO2 and the adiabatic flame temperature are fixed for a given refuse, the equation predicts
a linear relationship between temperature and C02 for a given refuse.
130
-------
APPENDIX E
CHARACTERIZATION OF THE BED-BURNING PROCESS
The overall objective of developing a model of the processes occurring within the
refuse bed is to determine the optimum distribution of both overfire and underfire. Previous
attempts at describing the bed burning have approached the burning as a combination of
several processes. The resulting models have been simplistic, providing little more than a
qualitative description of the refuse burning process. The difficulty does not lie in under-
standing the physical processes that are occurring, but rather in mathematically combining
the processes into a single coherent model.
The approach taken here is somewhat different. In the mixing study several phenom-
ena which could occur and control the bed-burning mechanism were identified. The
experimental data on burning obtained in this program allow these phenomena to be placed
in perspective so that conclusions regarding design relations can be drawn.
Following a series of short tests over each of the three grates of the Newton
incinerator, we selected the first three sample ports for complete bed tests. The data from
these tests are given in Table E-l. These two tests and the tests made in the breech form the
basis for conformation of the bed model presented here.
DESCRIPTION OF THE BURNING PROCESS
Several processes occur within the fuel bed, including drying, ignition, gasification,
combustion, and char burnout. Earlier qualitative models for fuel bed combustion identified
separate drying, gasification, and char-burning regions defined so that a single process was
occurring in each region. However, these operations actually occur simultaneously because
of the heterogeneous character of refuse.
The parameters controlling the burning bed are not well defined. Certainly one of the
more important parameters is air availability which is dependent upon how much and from
what region air is introduced into the furnace. Most municipal incinerators are designed to
introduce the primary combustion air from beneath the refuse bed, i.e., underfire air.
Consider a fuel bed supplied exclusively with underfire air. Oxygen will be consumed
as the underfire air passes through the bed so that, for underfire air rates less than that
required for complete combustion, there will occur a point within the bed where the
concentration of oxygen becomes zero. In a recent controlled study of incineration the
existence of this point was verified.*
131
-------
Table E-1. TEST RESULTS OVER BED1
Sample Port
Bed Test 1
1
Average
C02
02
CO
H2
HC
H2O2
H20/C02
C0/C02
Kwg
Temperature
Bed Test 2
C02
02
CO
H2
HC
H20
H20/C02
C0/C02
Kwg
Temperature
5.93
13.4
3.2
1.72
1.16
11.51
1.94
0.54
3.61
1835
13.3
6.0
3.35
2.7
1.94
12.9
0.97
0.25
1.20
1787
1.31
18.9
0.26
0.21
0.01
2.15
1.64
0.20
2.03
1905
6.4
14.2
1.00
0.77
0.34
19.4
3.03
0.16
3.94
1758
5.8
14.6
0.13
0.09
0.03
8.84
1.52
0.02
2.2
1928
12.0
8.6
0.26
0.19
0.39
26.4
1.70
0.02
2.33
1811
4.35
15.63
1.20
0.67
0.40
7.50
1.72
0.28
3.1
1889
10.6
9.6
1.54
1.22
0.89
17.57
1.66
0.15
2.09
1785
1. Complete sets of data are limited due to experimental difficulties directly over the bed.
2. Calculated from national balance.
In our tests, this was not the case. In both tests directly over the bed, the oxygen
concentration was on the order of 10%.
The presence of air in the bed offgas could be the result of any one of the three
mechanisms shown in Figure E-1.
132
-------
ENTRAPMENT
CHANNELING
EXCESS AIR
OVERFIRE
AIR
OVERFIRE
AIR
W////A
REFUSE /
V,
UNDERFIRE AIR
UNDERFIRE AIR
UNDERFIRE AIR
Figure E1 Oxygen breakthrough mechanisms
Excess Air
When the refuse bed burns to a point where the kinetics for the solid combustion
become very slow, then air can pass through the bed at a faster rate than that at which it can
be consumed. When the amount of excess air is small compared to the amount of air that is
consumed, the effect is of little consequence, since the air breaking through the bed will be
well mixed with the partial combustion products, and complete burnout of the complete
emissions will be eminent. However, when the fraction of oxygen breakthrough is much
larger than the amount of oxygen that can be consumed by the reaction, the heat liberated
by combustion will not be sufficient to raise the temperature of the gas stream and
quenching will occur. The latter situation is of particular importance in the final stages of
refuse burnout in which the refuse approaches a fuel-limiting condition. In the latter
fuel-limiting case, the gases are uniformly mixed and the equilibrium is controlled by
oxidation. In the case of both channeling and air entrainment in which non-uniform mixing
exists, the water gas shift equilibrium will apply to those portions of the gas which contain
combustibles but no oxygen. The non-uniformity of mixing is not unreasonable when
considering the mechanism by which the oxygen breakthrough occurs.
Air Entrainment
Several recent studies have shown that air can be drawn in from the overfire region and
entrained into the gases emanating from the refuse bed. In such a case, the bed-burning
process must be considered as a combination of both an underfire and an overfire mecha-
nism in which the zero-oxygen planes by each mechanism occur within the bed.
133
-------
Channeling
Because of the lack of a completely random packing of the refuse, channels, or "blow
holes," may develop. The "blow holes" allow the air to bypass the refuse completely. This
phenomenon is a common occurrence in packed bed equipment.
Each one of these three mechanisms could control the burning. In the case of the tests
at Newton, the overfire entrainment mechanism was believed to be the controlling factor.
This is best illustrated by considering the magnitude of local burning rates.
BURNING RATES
The overall burning rate can be determined from the amount of oxygen consumed,
from the amount ofCO2,CO, and hydrocarbons produced, or from the amount of heat
liberated. Table E-2 summarizes the overall burning rates derived from both carbon and
energy balances. The averages 54 and 67 lb/hr-ft2 of grate area based upon carbon and
energy balances, respectively, are close to the value of 60 lb/hr-ft2 often used as a design
value. Several assumptions on bed density, ash content, moisture content, and heating value
were introduced in the calculations. Errors in these assumed values may account for the
difference in burning rate obtained by the carbon and energy balances.
The fraction of total air consumed which could have been introduced underfire is also
given in Table E-2. Only one third of the total consumed air could have originated as
underfire air. In this case, the computed burning rates are 1.7 to 2.1 times that corre-
sponding to stoichiometric combustion of all of the underfire air.
The overfire entrainment mechanism is the only one which would result in burning
rates higher than the burning rate equivalent to complete combustion of the underfire air.
The results presented in Table E-2 also indicate that the overall burning rate is not
drastically affected by the underfire air rate, but remains relatively constant at about 60
lb/hr-ft2 of grate area in agreement with empirical design rules used for years in designing
burning surfaces. The rule holds because the total amount of air tends to remain constant.
Reductions in underfire air are offset by increases in overfire air entrainment.
OFFGAS COMPOSITION
To apply the concepts of "unmixedness" and "eddy decay," it is necessary to know
the initial composition of gases entering the turbulent mixing (overfire) region. From basic
research in kinetics of combustion reactions, we know that oxygen and hydrogen or carbon
monoxide react on the order of microseconds. For all practical purposes they cannot
co-exist. However, the data given in Table E-l show over 3% CO at times when oxygen is
over 10%. One reasonable explanation for this co-existence is that the offgases are relatively
unmixed so that the oxygen is not in intimate contact with the combustibles.
134
-------
Table E-2. SUMMARY OF BURNING RATES AND UNDERFIRE AIR CONSUMPTION
Burning Rate
(Ib/hr-ft2)
% of Consumed Air
Run No. * Heat Balance Mass Balance Introduced Underfire
BL 22/1
J 22/1
BL 23/1
JET 23/1
BL 23/2
JET 23/2
BL 24/1
JET 24/1
JET 24/2
BL 24/2
BL 25/1
JET 25/1
JET 25/2
BL 25/2
BL 31/1
JET 31/1
JET 31/2
BL 32/1
JET 32/1
JET 32/2
BL 33/1
JET 33/1
JET 33/2
BL 34/1
JET 34/1
JET 34/2
JET 34/3
("Total Heat ~|
[3900 x 336 J
67.9
65.6
70.2
55.3
66.3
59.0
68.9
75.1
62.4
70.7
79.5
75.8
58.6
56.4
75.3
61.5
57.7
77.1
77.3
74.2
67.4
64.8
71.4
67.6
62.0
60.8
61.6
rCQ2 x 23.21
[ (0.51) 336 J
Run conditions are given in Appendix A
53.3 35.5
56.5 31.3
50.1 41.5
41.5 40.6
43.7 39.3
40.5 38.5
43.8 35.7
59.8 29.6
51.4 30.6
56.4 30.6
70.9 35.7
60.0 31.5
53.3 34.9
46.6 34.9
69.2 27.04
50.8 27.04
47.6 27.04
i
59.7 28.9
66.7 30.4
53.5 28.5
56.0 29.6
54.9 29.2
57.9 29.2
55.6 33.2
49.5 30.04
49.2 23.06
53.9 30.04
135
-------
This being the case, the characterization of the offgas composition must include the
following factors:
1 ) Composition of combustible fraction,
2) Fraction of combustibles vs. fraction of air, and
3) Degree of unmixedness between two fractions.
These are discussed separately.
COMPOSITION OF COMBUSTIBLE FRACTION
The data gathered from earlier measurements* of bed offgas compositions suggest that
once the oxygen is depleted, a burning refuse bed acts as a gasifier, producing CO, CO2 , H2 ,
and H2 O with minor amounts of hydrocarbons. This must occur at some place within the
bed in order to have combustibles present in the offgas. The controlling equilibrium for this
portion of the offgas is between these gasifier products.
Tentative bed-burning models have been developed*** on the assumption that the
water gas shift reaction controls the gasification equilibrium. The experimental basis
supporting the postulated equilibrium was previously limited to data taken by Kaiser t on
the Oceanside (N.J..) incinerator and data taken at M.I.T. on synthetic refusett. The data
collected in this program also support this hypothesis.
The composition of the combustible fraction can be determined as a function of
CO/CO2 . The calculation scheme is as follows:
For the reaction:
CO2 + H2 -» CO + H2 O, (water gas shift) (E-l )
the equilibrium constant is defined as:
(H,0)(CO)
(E-2>
From the material balance:
CH04(H2O)06 -0.7H2O+ l.lOj -»CO2 + 1.5 H2O, (E-3)
136
-------
the equilibrium concentration can be calculated as follows:
Let: C0/C02 = x, then (H2 O) = ( J (H2 ) (E-4)
Choosing as a basis one mole of CO2 corresponding to x moles of CO, then the total
moles of carbon is (1 + x). From the material balance, the total hydrogen (in the form of
H2) is determined to be 1.5 (1 + x), since the ratio H2/C equals 1.5. Using the equilibrium
relation above and noting the sum of H2O and H2 equals the total hydrogen yields
1.5(1 + x); then the relative proportions of H2O and H2 can also be calculated. Finally, the
amount of N2 present is 3.76 times the amount of oxygen consumed. The latter is equal to
the total oxygen present (CO2 + 1/2 CO+ 1/2 H2) minus the amount of O2 originally
contained in the initial (1 + x) moles of refuse, i.e., 0.65 (1 + x). These proportions are
tabulated below.
Relative Proportion of Combustion Products
Product Mols
C02 1
CO x
1.5 (1 +x)K
H2O
H 1.5 (1 + x)x
I / 1.2K_ -0.3x
N2 3/76 l+
-------
u>
00
o
(N
D
D
D
Q
m ,
D
Q
[
a
D
a
a a
BD
3 D
Z
AA
A
3
D
n° a
A
O
A^b^
©A A
O
o
o
OQ
) GP AA Qj>>
C
O
D
o
0
750
O ZONE 1
A ZONE 2
D ZONE 3
1000
1250 1500
TEMPERATURE, °F
1750
2000
Figure E2 Plot of H2 O/CO2 versus temperature
-------
programs. The assumption is based on the concept that the mechanism for
forming CO and CO2 in the bed of an incinerator is the same as that for
gasification of other organic combustibles. This is approximately valid before
combustible burnout begins in the overfire region.
2. The ratio can be determined by actual sampling over the bed. The sampling
is extremely difficult and the analysis must always be suspect.
In both cases, the ratio is only a crude estimate, but the sophistication required to
predict the exact carbon ratio is neither currently available nor particularly warranted.
The analysis in Section V was based upon the CO/CO2 ratio as a characterization of
the extent to which the combustion reaction has gone to completion. From the summary of
the bed data in Table E-l an empirically determined ratio ranges from 0.02 to 0.6 with the
average being approximately 0.21. These values can be compared reasonably well with
values obtained for coal gasification or for pyrolysis of refuse. These values are probably
minimum because some overfire mixing has occurred before the gases from the bed can be
sampled in the sample probe. Thus, a certain amount of reaction would be expected within
the probe itself, reducing the amount of CO and correspondingly increasing the amount of
CO2 measured in each sample. Hence, the empirically determined CO/CO2 ratio is not
inconsistent with the types of initial ratios necessary to develop a condition in which
quenching would dominate (Appendix C).
A second tenet of the equilibrium model is that combustion is relatively uniform
throughout the bed, i.e., drying, gasification, combustion, burnout, etc, occur simultane-
ously. Although this is not rigorously true, the errors introduced by such an assumption are
no greater than those resulting from analytical measurements.
Verification of these two points: the equilibrium of gases and the uniformity through-
out the furnace is given below.
Two independent checks were made of the water-gas shift equilibrium. For the data
collected in the breech during the baseline tests, the water-gas shift constant is shown in
Figure E-3. With respect to these data, two things are noteworthy. First of all, the ratio
H2 O/CO2 remains relatively constant so that in the breech area the ratio CO/H2 is expected
to remain constant as well. The large amounts of water and carbon dioxide present act like a
flywheel, so that large changes in the CO and H2 concentrations are not at all seen as similar
changes in water or CO2. For refuse, the ratio H2O/CO2 is equal to 1.5 so that the water-gas
shift constant could be represented by the equation:
Kwg = 1.5CO/H2.
The fact that the calculated value of Kw g is even close to the theoretical value is a very
encouraging support of the water-gas equilibrium theory. The jet data were not used for this
because of the difficulty detecting H2 with the modified probe.
139
-------
15
10
o>
9
8
7
4
3
2
1
0
700
AA
KEY:
= (CO)(H20)
Wg " (CO,KH,)
^2/^2
o = ZONE I
n = ZONE II
A = ZONE III
Kwg (THEORETICAL)
1000
1250
1500
TEMP., °F
a
D
Figure E3 Water - gas shift constant from breech data
2000
140
-------
Additional confirmation- of the equilibrium can be seen from data taken directly over
the drying grate. In these tests in which the concentration of CO and H2 is sufficiently high
to obtain good analytical measurements, the experimental value for Kw g is remarkably close
to the theoretical value. A summary of the results of these tests was given in Table E-l. It is
also interesting to note that the ratio of hydrogen to carbon is consistent with the H2O/CO2
ratio of 1.5.
The condition of uniformity is also easy to substantiate by considering a plot of
H2O/CO2 as shown in Figure E-2. If, as the earlier models proposed, the drying zone and
the pyrolysis zone were distinct from one another, then one would expect to see a reduction
in the ratio of H2O/CO2 as the refuse travelled through the incinerator. For example, refuse
on the first grate would have only a small amount of combustion or pyrolysis and a large
amount of drying so that the ratio of H2O/CO2 would be larger than normal, reflecting the
large amounts of water resulting from the drying and small amounts of CO2 from combus-
tion. On the other hand, the gases over the burnout grate should be completely dried and
pyrolyzed so that very small amounts of hydrogen would be expected and the ratio of
H2O/CO2 would be correspondingly small. In contradiction to this, the data indicate that
there is no trend in H2 O/CO2, but rather that the ratio tends to be constant at approxi-
mately 1.5. The 1.5 value is consistent with the ratio of hydrogen to carbon for a refuse
with 25% moisture (by weight). This tendency was observed both in the breech and over the
bed and cannot, therefore, be attributed to overall mixing before the sample was taken.
Contrary to expectations, the H2O/CO2 ratio tends to be higher over the third grate
than it does over the other two. We do not attribute this to any characteristic of the
refuse-burning process, but rather to the fact that the hot ash dropping into the water-
quench hopper evaporates a sufficient amount of water to cause a shift in the H2 O/CO2
ratio over the third grate, an effect that would be witnessed in zone 3 because of the
stratification of temperature layers in the breech of the furnace. Indeed, a material balance
about the third grate indicates that the water introduced by this evaporation mechanism
would be sufficiently large for such an effect to be observed.
DETERMINATION OF OXYGEN FRACTION
Significant amounts of oxygen were present in most of the samples taken above the
fuel bed. Since, at furnace temperatures, oxygen does not coexist with supra-equilibrium
concentrations of CO and hydrogen for more than a few milliseconds, the detection of both
oxygen and fuel components in the samples is indicative of intermittent sampling from
fuel-rich and oxygen-rich eddies. The oxygen observed above the bed probably originates in
the underfire air which passes unreacted through blowholes or overfire air carried down by
buoyancy forces. Table E-3 presents estimates of the amount of oxygen present in the gas
samples expressed both as a fraction of the stoichiometric air requirements and a fraction of
141
-------
the total air in the sample (determined from a nitrogen balance). With the exception of one
sample, the available oxygen exceeded the stoichiometric oxygen requirements. The role of
overfire air jets for the present conditions therefore should be principally to provide mixing,
and caution should be exercised that the jets do not quench the combustion reactions.
Table E-3. AVAILABLE OXYGEN IN SAMPLES WITHDRAWN FROM ABOVE THE FIRST GRATE
Available O2 Unreacted 02 Unreacted O2
Bed Test 1
Probe Position 02 Required to Complete Combustion 02 Supplied Total Offgas
1 2.82 0.58
2 74 0.89 0.90 > 0.75
3 86 0.69
Bed Test 2
Probe Position
1 , 0.875 0.31
2 , 9.1 0.69 0.68 > 0.46
3 8.2 0.41
A peculiarity in the data resulting from the design of the incinerator merits comment.
The unreacted fraction in Table E-3 peaks at position 2, which was directly below one of
the roof air jets. An attempt made to close the damper to the roof jets between bed tests 1
and 2 resulted in a reduction of the unreacted air. This tended to support the hypothesis
that unreacted overfire air was entrained in the sample. Because of the presence of overfire
air, it is not possible to determine, on the basis of these measurements alone, the fraction of
the underfire air which is unreacted. The bed-burning data suggests that most of the
unreacted oxygen originates in the overfire jets.
DEGREE OF MIXING
The data given in Table E-l could be used to calculate an initial unmixedness factor
according to the procedure outlined in Appendix B. In fact, the value for $0 has already
been given in Table E- . The corresponding values for (A/C^)O are shown in Table E-4.
Values given for position 3, near the end of the first grate, are similar to the value for
(v/c^o calculated in Table E-3. One might be tempted to rationalize, explaining that
position 3 is closest to the center of the furnace, particularly to the first burning, and is
therefore more representative of the bulk of combustion gases found in zone I of the
breech. The data from zone I were used to calculate the values given in Table E-3.
Unfortunately, the data are too scattered and inconsistent for such a rationalization; no
conclusions are apparent from the data.
142
-------
Table E-4. INITIAL UNMIXEDNESS FACTORS
C0/C02
0.25
0.811
0.462
0.300
0.50
0.611
0.409
0.249
1.0
0.365
0.340
0.185
Position
1 2.3
2 0.60
3 0.58
Average 0.524 0.423 0.297
Bed 2
Position
1 - 2.0 1.273 2.982 4,951
2 1.14 0.556 0.457 0.332
3 1.18 0.286 0.184 0.061
0.705 1.208 1.781
143
-------
APPENDIX F
DERIVATION OF THE UNMIXEDNESS FACTOR
The following derivation is a summary of the development of the "unmixedness
factor" concept based upon Reference . The variable, 3>, the ratio of oxygen in the sample
to the oxygen required to complete combustion of the sample, can be calculated as follows:
If y is the mole fraction of offgas in the offgas/air mixture, then the mole fraction of
air is (1 y). Also, since 4.76 n moles of air react with 1 mole of offgas, then the
amount of offgas required to react with (1 - y) moles of air is [(1 - y)] /(4.76 n02 ). The
remaining offgas after the available oxygen has been consumed is equal to the initial moles
(y) minus the moles consumed, or in other words:
4.76
Since
Excess offgas = y - - (F-l)
L 4.76 -n02 J
/
\
(F'2)
excess offgas = a (y - ys) (F-3)
where
a = constant (1 +4.76 -n02)/(4.76 -n^). (F-4)
The oxygen required to complete combustion is n0 times the amount of offgas, so that:
Ot req = (C - Cs)
Likewise, in a sample where y is less than ys, the air concentration is still (1 y) Ct,
but in this case oxygen remains unreacted, i.e., the limiting reactant is offgas. In this case,
y Ct moles of offgas require (4.76 n0,)' (y ' Ct) moles of air. The remaining air is:
Excess air = (1 - y) Ct - (4.76 n0 a ) (y Ct)
= (l+4.76-n02)(y-ys)-Ct (F-S)
145
-------
and
n+4.76-n0.n
(C-C,) = a'(C-C,) (F-6)
Note that Eq. (F-6) is exactly the same as Eq. (F-4), except that it applies to the region
where y>ys, while Eq. (F-4) applies to the region where y/7rn" e ~" +erf(n) 1C
$ = *- ; n = ^= ^ (F-9)
1 + J e + erf(n)
A plot of 4> vs is given as Figure 3 in Section V.
146
-------
Length
Volume
Mass
APPENDIX G
CONVERSION FACTORS FOR SI UNITS
1 ft. = 0.3048 m
1 in. = 0.0254
1 cu. ft. = 0.02832 m3
1 Ib = 0.4536 kg
lib/ft3 = 16.017 kg/m3
llb/hr = 0.126xlQ-3 kg/sec
1 Ib/hr-ft2= 1.355 x 10"3 kg/m2-sec
1 ton/day = 0.0105 kg/sec
1 ton/hr = 0.438 x 10~3 kg/sec
Flow Rate
1 cfm = 0.472 xlO'3 m3/sec
1 ft/hr = 0.0847 x 10~3 m/sec
1 ft/sec = 0.3048 m/sec
Energy
1 Btu = 0.252 kcal
1 Btu/lb = 0.555 kcal/kg
1 Btu/lb°F = 1.0 kcal/kg °C
1 Btu/hr = 0.070 x 10~3 kcal/sec
1 Btu/hr ft2 = 0.754 x 10~3 kcal/m2 sec
1 Btu/hr ft3 = 2.47 x 10~3 kcal/m3 sec
Temperature
1°F = 0.555°C
Temp (°F) = 0.555 Temp (°C) + 32°F
or 1.8 [Temp (°F) - 32°F] = Temp (°C)
147
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-75-016
2.
3. RECIPIENT'S ACCESSION>NO.
4. TITLf AND SUBTITLE
Incinerator Overfire Mixing Demonstration
5. REPORT DATE
August 1975
6. PERFORMING ORGANIZATION CODE
7. AUTHORISL
T.J.Lamb, R.H.Stephens, C.M.Mohr, P.L.
Levins, L.K.Fox, and A. F.Sarofim
8. PERFORMING ORGANIZATION REPORT NO
ADL-73722
I. PERFORMING ORGANIZATION NAME AND ADDRESS
Arthur D. Little, Inc.
20 Acorn Park
Cambridge, MA 02140
10. PROGRAM ELEMENT NO.
1AB015; ROAP 21AUZ-015
11. CONTRACT/GRANT NO.
68-02-0204
12. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
EPA, Industrial Environmental Research Laboratory,
Research Triangle Park, NC 27711 and Municipal
Environmental Research Laboratory, Cincinnati,
Ohio 45268
Final: 1971 - 1974
14. SPONSORING AGENCY CODE
IS. SUPPLEMENTARY NOTES
16. ABSTRACT
The report gives test data and conclusions from a testing program of an existing
municipal incinerator. Measurements were made in the furnace breech, directly
over the refuse bed, and in the furnace stack. Generalized models were developed to
describe bed burning, furnace flow behavior, and overfire mixing phenomena. Design
guidelines were developed for operating variables, such as underfire air distribution
and feed rate, and design features, such as mixing jet placement and furnace config-
uration. Jets were shown to be effective in reducing combustible emissions through
temperature control. They were only marginally effective in inducing increased
turbulence within the furnace. Analysis of test results and rationale for the conclu-
sions drawn are discussed thoroughly.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COS AT I Field/Group
Air Pollution
Incinerators
Refuse Disposal
Combustion
Air Pollution Control
Stationary Sources
Overfire Mixing
13 B
21B
8. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
162
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
149
------- |